content recommendations case study

• Resources > 
• Netflix Content Recommendation System – P...

Netflix Content Recommendation System – Product Analytics Case Study

Netflix, the global streaming giant, owes much of its success to its sophisticated content recommendation system.

This system not only enhances user experience by providing personalized content but also drives viewer engagement and retention.

This case study delves into the mechanisms, algorithms, technical details, and impacts of Netflix’s recommendation system, providing insights into how it works and why it is so effective.

Netflix started as a DVD rental service in 1997, but it pivoted to streaming in 2007. As its library grew, so did the challenge of helping users find content they would enjoy. With millions of subscribers and thousands of titles, Netflix needed a system that could personalize recommendations to keep viewers engaged.

How the Recommendation System Works

Data Collection

• Viewing history: What a user watches, including the duration and frequency.
• Ratings: User ratings for watched content.
• Search queries: What users search for on the platform.
• Browsing behavior: Navigation patterns, click-through rates, and scroll depths.
• Interaction timestamps: When content is watched (time of day, day of the week).
• Genre, sub-genre, and micro-genres.
• Cast, director, and crew information.
• Release year and content duration.
• User-generated tags and descriptions.
• Detailed synopsis and themes.

Algorithms and Techniques

• User-Based Collaborative Filtering : Recommendations are made based on the preferences of similar users. If User A and User B have similar viewing habits, content enjoyed by User A is likely to be recommended to User B.
• Item-Based Collaborative Filtering : This method looks at the relationship between items. If two items are often watched together, they are likely to be recommended in tandem.
• Matrix Factorization : Techniques like Singular Value Decomposition (SVD) break down large matrices (user-item interactions) into lower-dimensional spaces to uncover latent factors that influence preferences.
• This method recommends content similar to what a user has watched in the past. For instance, if a user enjoys science fiction movies, the system suggests more titles in the same genre.
• Natural Language Processing (NLP) : Used to analyze and understand the content metadata and user reviews to extract meaningful features.
• Convolutional Neural Networks (CNNs) : Used for image processing, helping to analyze and recommend based on video frames and thumbnails.
• Recurrent Neural Networks (RNNs) and Long Short-Term Memory Networks (LSTMs) : Effective in capturing temporal patterns in viewing behavior, making predictions based on sequential data.
• Variational Autoencoders (VAEs) and Generative Adversarial Networks (GANs) : Used for generating new content recommendations by learning complex distributions of user preferences.
• Multi-armed bandit algorithms are employed for exploration-exploitation trade-offs, balancing between recommending popular content and exploring less-known titles that might interest the user.
• Contextual Bandits : Incorporate context (e.g., time of day, user device) to make more informed recommendations.

Personalization Features

• Homepage Rows : Rows on the Netflix homepage are personalized for each user, highlighting different genres, trending now, or continue watching. The order and content of these rows are dynamically adjusted based on user behavior.
• Artwork Personalization : The thumbnail images for each title are also personalized. Different users see different images for the same title based on what aspects might appeal to them (e.g., an action scene vs. a romantic moment).

Contextual Recommendations

• Recommendations are not only personalized but also contextualized. For example, weekend recommendations might differ from weekday recommendations based on viewing patterns.

Technical Architecture

1. Data Pipeline

• Data Ingestion : Massive amounts of data are ingested from various sources (user interactions, content metadata) in real-time.
• Data Storage : Data is stored in distributed systems like Apache Cassandra and Amazon S3 to handle scale and ensure durability.
• Data Processing : Apache Kafka is used for real-time data streaming, while Apache Spark and Flink handle batch and stream processing.

2. Model Training and Deployment

• Feature Engineering : Complex features are engineered from raw data, such as user embeddings, content embeddings, and temporal features.
• Model Training : Models are trained on distributed computing frameworks like TensorFlow, PyTorch, and Horovod, leveraging GPU clusters.
• Model Deployment : Trained models are deployed using containerized environments (Docker, Kubernetes) to ensure scalability and reliability.
• Continuous Learning : The system continuously learns from new data. Online learning algorithms and incremental training approaches are used to keep models updated.

3. Scalability and Performance

• Netflix employs a microservices architecture to handle different aspects of the recommendation system. This ensures modularity and ease of maintenance.
• Load balancing and caching strategies are implemented to handle high traffic volumes and ensure low latency in serving recommendations.

Evaluation and Testing

Netflix employs extensive A/B testing to evaluate the effectiveness of its recommendation algorithms. Different versions of the recommendation system are tested with various user segments to determine which version improves user engagement and satisfaction the most. Key metrics include:

• Click-Through Rate (CTR) : Measures how often users click on recommended content.
• View Duration : Tracks how long users watch the recommended content.
• User Retention : Analyzes the impact of recommendations on subscription renewals.

Impacts on Business

1. Increased User Engagement

• Personalized recommendations keep users watching longer, reducing churn and increasing subscription longevity.
• Netflix reports that personalized recommendations drive over 80% of the content watched on the platform.

2. Content Discovery

• The recommendation system helps surface lesser-known titles, ensuring that even niche content finds its audience. This helps in maximizing the return on investment for all content acquired or produced.

3. Operational Efficiency

• By accurately predicting user preferences, Netflix can make better-informed decisions about which new content to acquire or produce, optimizing content spend.

4. Customer Satisfaction

• Personalized experiences lead to higher customer satisfaction, reflected in positive reviews and word-of-mouth referrals.

Challenges and Future Directions

• Data Privacy
• Collecting and using personal data responsibly is crucial. Netflix ensures user data is anonymized and used ethically.

2. Diverse Preferences

• Balancing the diverse tastes of a global audience remains a challenge. Continuous improvements in localization and personalized algorithms are needed.

3. Algorithmic Bias

• Ensuring that recommendation algorithms do not reinforce existing biases is essential. Netflix works to maintain fairness and diversity in recommendations.

4. Real-Time Personalization

• Enhancing real-time personalization to provide instant recommendations based on immediate user actions and changing contexts.

Netflix’s content recommendation system is a cornerstone of its success, blending sophisticated algorithms with vast amounts of data to deliver a personalized viewing experience.

The system’s ability to keep users engaged and help them discover new content has been pivotal in Netflix’s rise to become a leading entertainment service globally.

The ongoing advancements in AI and machine learning ensure that Netflix remains at the forefront of personalized content delivery, continually enhancing user experience and business outcomes.

• Netflix Tech Blog. “Netflix Recommendations: Beyond the 5 stars (Part 1).” Medium, 6 April 2017. Netflix Tech Blog .
• Amatriain, Xavier, and Justin Basilico. “Netflix Recommendations: Beyond the 5 stars (Part 2).” Medium, 20 June 2012. Netflix Tech Blog .
• Gomez-Uribe, Carlos A., and Neil Hunt. “The Netflix Recommender System: Algorithms, Business Value, and Innovation.” ACM Transactions on Management Information Systems (TMIS), vol. 6, no. 4, 2015, pp. 13-20.
• “How Netflix’s Recommendations System Works.” Netflix, 2021. Netflix Help Center .
• “Netflix Personalization Explained.” Netflix Research, 2020. Netflix Research .
• Zhou, Y., Wilkinson, D., Schreiber, R., & Pan, R. “Large-Scale Parallel Collaborative Filtering for the Netflix Prize.” Lecture Notes in Computer Science, 2008.
• Sedhain, S., Menon, A. K., Sanner, S., & Xie, L. “Autorec: Autoencoders Meet Collaborative Filtering.” Proceedings of the 24th International Conference on World Wide Web, 2015.
• Koren, Y., Bell, R., & Volinsky, C. “Matrix Factorization Techniques for Recommender Systems.” Computer, vol. 42, no. 8, 2009, pp. 30-37.

Learn Practical Product Management from Top Mentors & Leaders

🔓 Unlock This & Grab our Top Product Resources for Free

⚡️ Join our "value-packed" weekly product newsletter

• Design for Business
• Most Recent
• Presentations
• Infographics
• Data Visualizations
• Forms and Surveys
• Video & Animation
• Case Studies
• Digital Marketing
• Design Inspiration
• Visual Thinking
• Product Updates
• Visme Webinars
• Artificial Intelligence

15 Real-Life Case Study Examples & Best Practices

Written by: Oghale Olori

Case studies are more than just success stories.

They are powerful tools that demonstrate the practical value of your product or service. Case studies help attract attention to your products, build trust with potential customers and ultimately drive sales.

It’s no wonder that 73% of successful content marketers utilize case studies as part of their content strategy. Plus, buyers spend 54% of their time reviewing case studies before they make a buying decision.

To ensure you’re making the most of your case studies, we’ve put together 15 real-life case study examples to inspire you. These examples span a variety of industries and formats. We’ve also included best practices, design tips and templates to inspire you.

Let’s dive in!

Table of Contents

What is a case study, 15 real-life case study examples, sales case study examples, saas case study examples, product case study examples, marketing case study examples, business case study examples, case study faqs.

• A case study is a compelling narrative that showcases how your product or service has positively impacted a real business or individual. 
• Case studies delve into your customer's challenges, how your solution addressed them and the quantifiable results they achieved.
• Your case study should have an attention-grabbing headline, great visuals and a relevant call to action. Other key elements include an introduction, problems and result section.
• Visme provides easy-to-use tools, professionally designed templates and features for creating attractive and engaging case studies.

A case study is a real-life scenario where your company helped a person or business solve their unique challenges. It provides a detailed analysis of the positive outcomes achieved as a result of implementing your solution.

Case studies are an effective way to showcase the value of your product or service to potential customers without overt selling. By sharing how your company transformed a business, you can attract customers seeking similar solutions and results.

Case studies are not only about your company's capabilities; they are primarily about the benefits customers and clients have experienced from using your product.

Every great case study is made up of key elements. They are;

• Attention-grabbing headline: Write a compelling headline that grabs attention and tells your reader what the case study is about. For example, "How a CRM System Helped a B2B Company Increase Revenue by 225%.
• Introduction/Executive Summary: Include a brief overview of your case study, including your customer’s problem, the solution they implemented and the results they achieved.
• Problem/Challenge: Case studies with solutions offer a powerful way to connect with potential customers. In this section, explain how your product or service specifically addressed your customer's challenges.
• Solution: Explain how your product or service specifically addressed your customer's challenges.
• Results/Achievements : Give a detailed account of the positive impact of your product. Quantify the benefits achieved using metrics such as increased sales, improved efficiency, reduced costs or enhanced customer satisfaction.
• Graphics/Visuals: Include professional designs, high-quality photos and videos to make your case study more engaging and visually appealing.
• Quotes/Testimonials: Incorporate written or video quotes from your clients to boost your credibility.
• Relevant CTA: Insert a call to action (CTA) that encourages the reader to take action. For example, visiting your website or contacting you for more information. Your CTA can be a link to a landing page, a contact form or your social media handle and should be related to the product or service you highlighted in your case study.

Now that you understand what a case study is, let’s look at real-life case study examples. Among these, you'll find some simple case study examples that break down complex ideas into easily understandable solutions.

In this section, we’ll explore SaaS, marketing, sales, product and business case study examples with solutions. Take note of how these companies structured their case studies and included the key elements.

We’ve also included professionally designed case study templates to inspire you.

1. Georgia Tech Athletics Increase Season Ticket Sales by 80%

Georgia Tech Athletics, with its 8,000 football season ticket holders, sought for a way to increase efficiency and customer engagement.

Their initial sales process involved making multiple outbound phone calls per day with no real targeting or guidelines. Georgia Tech believed that targeting communications will enable them to reach more people in real time.

Salesloft improved Georgia Tech’s sales process with an inbound structure. This enabled sales reps to connect with their customers on a more targeted level. The use of dynamic fields and filters when importing lists ensured prospects received the right information, while communication with existing fans became faster with automation.

As a result, Georgia Tech Athletics recorded an 80% increase in season ticket sales as relationships with season ticket holders significantly improved. Employee engagement increased as employees became more energized to connect and communicate with fans.

Why Does This Case Study Work?

In this case study example , Salesloft utilized the key elements of a good case study. Their introduction gave an overview of their customers' challenges and the results they enjoyed after using them. After which they categorized the case study into three main sections: challenge, solution and result.

Salesloft utilized a case study video to increase engagement and invoke human connection.

Incorporating videos in your case study has a lot of benefits. Wyzol’s 2023 state of video marketing report showed a direct correlation between videos and an 87% increase in sales.

The beautiful thing is that creating videos for your case study doesn’t have to be daunting.

With an easy-to-use platform like Visme, you can create top-notch testimonial videos that will connect with your audience. Within the Visme editor, you can access over 1 million stock photos , video templates, animated graphics and more. These tools and resources will significantly improve the design and engagement of your case study.

Simplify content creation and brand management for your team

• Collaborate on designs , mockups and wireframes with your non-design colleagues
• Lock down your branding to maintain brand consistency throughout your designs
• Why start from scratch? Save time with 1000s of professional branded templates

Sign up. It’s free.

2. WeightWatchers Completely Revamped their Enterprise Sales Process with HubSpot

WeightWatchers, a 60-year-old wellness company, sought a CRM solution that increased the efficiency of their sales process. With their previous system, Weightwatchers had limited automation. They would copy-paste message templates from word documents or recreate one email for a batch of customers.

This required a huge effort from sales reps, account managers and leadership, as they were unable to track leads or pull customized reports for planning and growth.

WeightWatchers transformed their B2B sales strategy by leveraging HubSpot's robust marketing and sales workflows. They utilized HubSpot’s deal pipeline and automation features to streamline lead qualification. And the customized dashboard gave leadership valuable insights.

As a result, WeightWatchers generated seven figures in annual contract value and boosted recurring revenue. Hubspot’s impact resulted in 100% adoption across all sales, marketing, client success and operations teams.

Hubspot structured its case study into separate sections, demonstrating the specific benefits of their products to various aspects of the customer's business. Additionally, they integrated direct customer quotes in each section to boost credibility, resulting in a more compelling case study.

Getting insight from your customer about their challenges is one thing. But writing about their process and achievements in a concise and relatable way is another. If you find yourself constantly experiencing writer’s block, Visme’s AI writer is perfect for you.

Visme created this AI text generator tool to take your ideas and transform them into a great draft. So whether you need help writing your first draft or editing your final case study, Visme is ready for you.

3. Immi’s Ram Fam Helps to Drive Over $200k in Sales

Immi embarked on a mission to recreate healthier ramen recipes that were nutritious and delicious. After 2 years of tireless trials, Immi finally found the perfect ramen recipe. However, they envisioned a community of passionate ramen enthusiasts to fuel their business growth.

This vision propelled them to partner with Shopify Collabs. Shopify Collabs successfully cultivated and managed Immi’s Ramen community of ambassadors and creators.

As a result of their partnership, Immi’s community grew to more than 400 dedicated members, generating over $200,000 in total affiliate sales.

The power of data-driven headlines cannot be overemphasized. Chili Piper strategically incorporates quantifiable results in their headlines. This instantly sparks curiosity and interest in readers.

While not every customer success story may boast headline-grabbing figures, quantifying achievements in percentages is still effective. For example, you can highlight a 50% revenue increase with the implementation of your product.

Take a look at the beautiful case study template below. Just like in the example above, the figures in the headline instantly grab attention and entice your reader to click through.

Having a case study document is a key factor in boosting engagement. This makes it easy to promote your case study in multiple ways. With Visme, you can easily publish, download and share your case study with your customers in a variety of formats, including PDF, PPTX, JPG and more!

4. How WOW! is Saving Nearly 79% in Time and Cost With Visme

This case study discusses how Visme helped WOW! save time and money by providing user-friendly tools to create interactive and quality training materials for their employees. Find out what your team can do with Visme. Request a Demo

WOW!'s learning and development team creates high-quality training materials for new and existing employees. Previous tools and platforms they used had plain templates, little to no interactivity features, and limited flexibility—that is, until they discovered Visme.

Now, the learning and development team at WOW! use Visme to create engaging infographics, training videos, slide decks and other training materials.

This has directly reduced the company's turnover rate, saving them money spent on recruiting and training new employees. It has also saved them a significant amount of time, which they can now allocate to other important tasks.

Visme's customer testimonials spark an emotional connection with the reader, leaving a profound impact. Upon reading this case study, prospective customers will be blown away by the remarkable efficiency achieved by Visme's clients after switching from PowerPoint.

Visme’s interactivity feature was a game changer for WOW! and one of the primary reasons they chose Visme.

“Previously we were using PowerPoint, which is fine, but the interactivity you can get with Visme is so much more robust that we’ve all steered away from PowerPoint.” - Kendra, L&D team, Wow!

Visme’s interactive feature allowed them to animate their infographics, include clickable links on their PowerPoint designs and even embed polls and quizzes their employees could interact with.

By embedding the slide decks, infographics and other training materials WOW! created with Visme, potential customers get a taste of what they can create with the tool. This is much more effective than describing the features of Visme because it allows potential customers to see the tool in action.

To top it all off, this case study utilized relevant data and figures. For example, one part of the case study said, “In Visme, where Kendra’s team has access to hundreds of templates, a brand kit, and millions of design assets at their disposal, their team can create presentations in 80% less time.”

Who wouldn't want that?

Including relevant figures and graphics in your case study is a sure way to convince your potential customers why you’re a great fit for their brand. The case study template below is a great example of integrating relevant figures and data.

This colorful template begins with a captivating headline. But that is not the best part; this template extensively showcases the results their customer had using relevant figures.

The arrangement of the results makes it fun and attractive. Instead of just putting figures in a plain table, you can find interesting shapes in your Visme editor to take your case study to the next level.

5. Lyte Reduces Customer Churn To Just 3% With Hubspot CRM

While Lyte was redefining the ticketing industry, it had no definite CRM system . Lyte utilized 12–15 different SaaS solutions across various departments, which led to a lack of alignment between teams, duplication of work and overlapping tasks.

Customer data was spread across these platforms, making it difficult to effectively track their customer journey. As a result, their churn rate increased along with customer dissatisfaction.

Through Fuelius , Lyte founded and implemented Hubspot CRM. Lyte's productivity skyrocketed after incorporating Hubspot's all-in-one CRM tool. With improved efficiency, better teamwork and stronger client relationships, sales figures soared.

The case study title page and executive summary act as compelling entry points for both existing and potential customers. This overview provides a clear understanding of the case study and also strategically incorporates key details like the client's industry, location and relevant background information.

Having a good summary of your case study can prompt your readers to engage further. You can achieve this with a simple but effective case study one-pager that highlights your customer’s problems, process and achievements, just like this case study did in the beginning.

Moreover, you can easily distribute your case study one-pager and use it as a lead magnet to draw prospective customers to your company.

Take a look at this case study one-pager template below.

This template includes key aspects of your case study, such as the introduction, key findings, conclusion and more, without overcrowding the page. The use of multiple shades of blue gives it a clean and dynamic layout.

Our favorite part of this template is where the age group is visualized.

With Visme’s data visualization tool , you can present your data in tables, graphs, progress bars, maps and so much more. All you need to do is choose your preferred data visualization widget, input or import your data and click enter!

6. How Workato Converts 75% of Their Qualified Leads

Workato wanted to improve their inbound leads and increase their conversion rate, which ranged from 40-55%.

At first, Workato searched for a simple scheduling tool. They soon discovered that they needed a tool that provided advanced routing capabilities based on zip code and other criteria. Luckily, they found and implemented Chili Piper.

As a result of implementing Chili Piper, Workato achieved a remarkable 75–80% conversion rate and improved show rates. This led to a substantial revenue boost, with a 10-15% increase in revenue attributed to Chili Piper's impact on lead conversion.

This case study example utilizes the power of video testimonials to drive the impact of their product.

Chili Piper incorporates screenshots and clips of their tool in use. This is a great strategy because it helps your viewers become familiar with how your product works, making onboarding new customers much easier.

In this case study example, we see the importance of efficient Workflow Management Systems (WMS). Without a WMS, you manually assign tasks to your team members and engage in multiple emails for regular updates on progress.

However, when crafting and designing your case study, you should prioritize having a good WMS.

Visme has an outstanding Workflow Management System feature that keeps you on top of all your projects and designs. This feature makes it much easier to assign roles, ensure accuracy across documents, and track progress and deadlines.

Visme’s WMS feature allows you to limit access to your entire document by assigning specific slides or pages to individual members of your team. At the end of the day, your team members are not overwhelmed or distracted by the whole document but can focus on their tasks.

7. Rush Order Helps Vogmask Scale-Up During a Pandemic

Vomask's reliance on third-party fulfillment companies became a challenge as demand for their masks grew. Seeking a reliable fulfillment partner, they found Rush Order and entrusted them with their entire inventory.

Vomask's partnership with Rush Order proved to be a lifesaver during the COVID-19 pandemic. Rush Order's agility, efficiency and commitment to customer satisfaction helped Vogmask navigate the unprecedented demand and maintain its reputation for quality and service.

Rush Order’s comprehensive support enabled Vogmask to scale up its order processing by a staggering 900% while maintaining a remarkable customer satisfaction rate of 92%.

Rush Order chose one event where their impact mattered the most to their customer and shared that story.

While pandemics don't happen every day, you can look through your customer’s journey and highlight a specific time or scenario where your product or service saved their business.

The story of Vogmask and Rush Order is compelling, but it simply is not enough. The case study format and design attract readers' attention and make them want to know more. Rush Order uses consistent colors throughout the case study, starting with the logo, bold square blocks, pictures, and even headers.

Take a look at this product case study template below.

Just like our example, this case study template utilizes bold colors and large squares to attract and maintain the reader’s attention. It provides enough room for you to write about your customers' backgrounds/introductions, challenges, goals and results.

The right combination of shapes and colors adds a level of professionalism to this case study template.

8. AMR Hair & Beauty leverages B2B functionality to boost sales by 200%

With limits on website customization, slow page loading and multiple website crashes during peak events, it wasn't long before AMR Hair & Beauty began looking for a new e-commerce solution.

Their existing platform lacked effective search and filtering options, a seamless checkout process and the data analytics capabilities needed for informed decision-making. This led to a significant number of abandoned carts.

Upon switching to Shopify Plus, AMR immediately saw improvements in page loading speed and average session duration. They added better search and filtering options for their wholesale customers and customized their checkout process.

Due to this, AMR witnessed a 200% increase in sales and a 77% rise in B2B average order value. AMR Hair & Beauty is now poised for further expansion and growth.

This case study example showcases the power of a concise and impactful narrative.

To make their case analysis more effective, Shopify focused on the most relevant aspects of the customer's journey. While there may have been other challenges the customer faced, they only included those that directly related to their solutions.

Take a look at this case study template below. It is perfect if you want to create a concise but effective case study. Without including unnecessary details, you can outline the challenges, solutions and results your customers experienced from using your product.

Don’t forget to include a strong CTA within your case study. By incorporating a link, sidebar pop-up or an exit pop-up into your case study, you can prompt your readers and prospective clients to connect with you.

9. How a Marketing Agency Uses Visme to Create Engaging Content With Infographics

SmartBox Dental , a marketing agency specializing in dental practices, sought ways to make dental advice more interesting and easier to read. However, they lacked the design skills to do so effectively.

Visme's wide range of templates and features made it easy for the team to create high-quality content quickly and efficiently. SmartBox Dental enjoyed creating infographics in as little as 10-15 minutes, compared to one hour before Visme was implemented.

By leveraging Visme, SmartBox Dental successfully transformed dental content into a more enjoyable and informative experience for their clients' patients. Therefore enhancing its reputation as a marketing partner that goes the extra mile to deliver value to its clients.

Visme creatively incorporates testimonials In this case study example.

By showcasing infographics and designs created by their clients, they leverage the power of social proof in a visually compelling way. This way, potential customers gain immediate insight into the creative possibilities Visme offers as a design tool.

This example effectively showcases a product's versatility and impact, and we can learn a lot about writing a case study from it. Instead of focusing on one tool or feature per customer, Visme took a more comprehensive approach.

Within each section of their case study, Visme explained how a particular tool or feature played a key role in solving the customer's challenges.

For example, this case study highlighted Visme’s collaboration tool . With Visme’s tool, the SmartBox Dental content team fostered teamwork, accountability and effective supervision.

Visme also achieved a versatile case study by including relevant quotes to showcase each tool or feature. Take a look at some examples;

Visme’s collaboration tool: “We really like the collaboration tool. Being able to see what a co-worker is working on and borrow their ideas or collaborate on a project to make sure we get the best end result really helps us out.”

Visme’s library of stock photos and animated characters: “I really love the images and the look those give to an infographic. I also really like the animated little guys and the animated pictures. That’s added a lot of fun to our designs.”

Visme’s interactivity feature: “You can add URLs and phone number links directly into the infographic so they can just click and call or go to another page on the website and I really like adding those hyperlinks in.”

You can ask your customers to talk about the different products or features that helped them achieve their business success and draw quotes from each one.

10. Jasper Grows Blog Organic Sessions 810% and Blog-Attributed User Signups 400X

Jasper, an AI writing tool, lacked a scalable content strategy to drive organic traffic and user growth. They needed help creating content that converted visitors into users. Especially when a looming domain migration threatened organic traffic.

To address these challenges, Jasper partnered with Omniscient Digital. Their goal was to turn their content into a growth channel and drive organic growth. Omniscient Digital developed a full content strategy for Jasper AI, which included a content audit, competitive analysis, and keyword discovery.

Through their collaboration, Jasper’s organic blog sessions increased by 810%, despite the domain migration. They also witnessed a 400X increase in blog-attributed signups. And more importantly, the content program contributed to over $4 million in annual recurring revenue.

The combination of storytelling and video testimonials within the case study example makes this a real winner. But there’s a twist to it. Omniscient segmented the video testimonials and placed them in different sections of the case study.

Video marketing , especially in case studies, works wonders. Research shows us that 42% of people prefer video testimonials because they show real customers with real success stories. So if you haven't thought of it before, incorporate video testimonials into your case study.

Take a look at this stunning video testimonial template. With its simple design, you can input the picture, name and quote of your customer within your case study in a fun and engaging way.

Try it yourself! Customize this template with your customer’s testimonial and add it to your case study!

11. How Meliá Became One of the Most Influential Hotel Chains on Social Media

Meliá Hotels needed help managing their growing social media customer service needs. Despite having over 500 social accounts, they lacked a unified response protocol and detailed reporting. This largely hindered efficiency and brand consistency.

Meliá partnered with Hootsuite to build an in-house social customer care team. Implementing Hootsuite's tools enabled Meliá to decrease response times from 24 hours to 12.4 hours while also leveraging smart automation.

In addition to that, Meliá resolved over 133,000 conversations, booking 330 inquiries per week through Hootsuite Inbox. They significantly improved brand consistency, response time and customer satisfaction.

The need for a good case study design cannot be over-emphasized.

As soon as anyone lands on this case study example, they are mesmerized by a beautiful case study design. This alone raises the interest of readers and keeps them engaged till the end.

If you’re currently saying to yourself, “ I can write great case studies, but I don’t have the time or skill to turn it into a beautiful document.” Say no more.

Visme’s amazing AI document generator can take your text and transform it into a stunning and professional document in minutes! Not only do you save time, but you also get inspired by the design.

With Visme’s document generator, you can create PDFs, case study presentations , infographics and more!

Take a look at this case study template below. Just like our case study example, it captures readers' attention with its beautiful design. Its dynamic blend of colors and fonts helps to segment each element of the case study beautifully.

12. Tea’s Me Cafe: Tamika Catchings is Brewing Glory

Tamika's journey began when she purchased Tea's Me Cafe in 2017, saving it from closure. She recognized the potential of the cafe as a community hub and hosted regular events centered on social issues and youth empowerment.

One of Tamika’s business goals was to automate her business. She sought to streamline business processes across various aspects of her business. One of the ways she achieves this goal is through Constant Contact.

Constant Contact became an integral part of Tamika's marketing strategy. They provided an automated and centralized platform for managing email newsletters, event registrations, social media scheduling and more.

This allowed Tamika and her team to collaborate efficiently and focus on engaging with their audience. They effectively utilized features like WooCommerce integration, text-to-join and the survey builder to grow their email list, segment their audience and gather valuable feedback.

The case study example utilizes the power of storytelling to form a connection with readers. Constant Contact takes a humble approach in this case study. They spotlight their customers' efforts as the reason for their achievements and growth, establishing trust and credibility.

This case study is also visually appealing, filled with high-quality photos of their customer. While this is a great way to foster originality, it can prove challenging if your customer sends you blurry or low-quality photos.

If you find yourself in that dilemma, you can use Visme’s AI image edit tool to touch up your photos. With Visme’s AI tool, you can remove unwanted backgrounds, erase unwanted objects, unblur low-quality pictures and upscale any photo without losing the quality.

Constant Contact offers its readers various formats to engage with their case study. Including an audio podcast and PDF.

In its PDF version, Constant Contact utilized its brand colors to create a stunning case study design.  With this, they increase brand awareness and, in turn, brand recognition with anyone who comes across their case study.

With Visme’s brand wizard tool , you can seamlessly incorporate your brand assets into any design or document you create. By inputting your URL, Visme’s AI integration will take note of your brand colors, brand fonts and more and create branded templates for you automatically.

You don't need to worry about spending hours customizing templates to fit your brand anymore. You can focus on writing amazing case studies that promote your company.

13. How Breakwater Kitchens Achieved a 7% Growth in Sales With Thryv

Breakwater Kitchens struggled with managing their business operations efficiently. They spent a lot of time on manual tasks, such as scheduling appointments and managing client communication. This made it difficult for them to grow their business and provide the best possible service to their customers.

David, the owner, discovered Thryv. With Thryv, Breakwater Kitchens was able to automate many of their manual tasks. Additionally, Thryv integrated social media management. This enabled Breakwater Kitchens to deliver a consistent brand message, captivate its audience and foster online growth.

As a result, Breakwater Kitchens achieved increased efficiency, reduced missed appointments and a 7% growth in sales.

This case study example uses a concise format and strong verbs, which make it easy for readers to absorb the information.

At the top of the case study, Thryv immediately builds trust by presenting their customer's complete profile, including their name, company details and website. This allows potential customers to verify the case study's legitimacy, making them more likely to believe in Thryv's services.

However, manually copying and pasting customer information across multiple pages of your case study can be time-consuming.

To save time and effort, you can utilize Visme's dynamic field feature . Dynamic fields automatically insert reusable information into your designs.  So you don’t have to type it out multiple times.

14. Zoom’s Creative Team Saves Over 4,000 Hours With Brandfolder

Zoom experienced rapid growth with the advent of remote work and the rise of the COVID-19 pandemic. Such growth called for agility and resilience to scale through.

At the time, Zoom’s assets were disorganized which made retrieving brand information a burden. Zoom’s creative manager spent no less than 10 hours per week finding and retrieving brand assets for internal teams.

Zoom needed a more sustainable approach to organizing and retrieving brand information and came across Brandfolder. Brandfolder simplified and accelerated Zoom’s email localization and webpage development. It also enhanced the creation and storage of Zoom virtual backgrounds.

With Brandfolder, Zoom now saves 4,000+ hours every year. The company also centralized its assets in Brandfolder, which allowed 6,800+ employees and 20-30 vendors to quickly access them.

Brandfolder infused its case study with compelling data and backed it up with verifiable sources. This data-driven approach boosts credibility and increases the impact of their story.

Bradfolder's case study goes the extra mile by providing a downloadable PDF version, making it convenient for readers to access the information on their own time. Their dedication to crafting stunning visuals is evident in every aspect of the project.

From the vibrant colors to the seamless navigation, everything has been meticulously designed to leave a lasting impression on the viewer. And with clickable links that make exploring the content a breeze, the user experience is guaranteed to be nothing short of exceptional.

The thing is, your case study presentation won’t always sit on your website. There are instances where you may need to do a case study presentation for clients, partners or potential investors.

Visme has a rich library of templates you can tap into. But if you’re racing against the clock, Visme’s AI presentation maker is your best ally.

15. How Cents of Style Made $1.7M+ in Affiliate Sales with LeadDyno

Cents of Style had a successful affiliate and influencer marketing strategy. However, their existing affiliate marketing platform was not intuitive, customizable or transparent enough to meet the needs of their influencers.

Cents of Styles needed an easy-to-use affiliate marketing platform that gave them more freedom to customize their program and implement a multi-tier commission program.

After exploring their options, Cents of Style decided on LeadDyno.

LeadDyno provided more flexibility, allowing them to customize commission rates and implement their multi-tier commission structure, switching from monthly to weekly payouts.

Also, integrations with PayPal made payments smoother And features like newsletters and leaderboards added to the platform's success by keeping things transparent and engaging.

As a result, Cents of Style witnessed an impressive $1.7 million in revenue from affiliate sales with a substantial increase in web sales by 80%.

LeadDyno strategically placed a compelling CTA in the middle of their case study layout, maximizing its impact. At this point, readers are already invested in the customer's story and may be considering implementing similar strategies.

A well-placed CTA offers them a direct path to learn more and take action.

LeadDyno also utilized the power of quotes to strengthen their case study. They didn't just embed these quotes seamlessly into the text; instead, they emphasized each one with distinct blocks.

Are you looking for an easier and quicker solution to create a case study and other business documents? Try Visme's AI designer ! This powerful tool allows you to generate complete documents, such as case studies, reports, whitepapers and more, just by providing text prompts. Simply explain your requirements to the tool, and it will produce the document for you, complete with text, images, design assets and more.

Still have more questions about case studies? Let's look at some frequently asked questions.

How to Write a Case Study?

• Choose a compelling story: Not all case studies are created equal. Pick one that is relevant to your target audience and demonstrates the specific benefits of your product or service.
• Outline your case study: Create a case study outline and highlight how you will structure your case study to include the introduction, problem, solution and achievements of your customer.
• Choose a case study template: After you outline your case study, choose a case study template . Visme has stunning templates that can inspire your case study design.
• Craft a compelling headline: Include figures or percentages that draw attention to your case study.
• Work on the first draft: Your case study should be easy to read and understand. Use clear and concise language and avoid jargon.
• Include high-quality visual aids: Visuals can help to make your case study more engaging and easier to read. Consider adding high-quality photos, screenshots or videos.
• Include a relevant CTA: Tell prospective customers how to reach you for questions or sign-ups.

What Are the Stages of a Case Study?

The stages of a case study are;

• Planning & Preparation: Highlight your goals for writing the case study. Plan the case study format, length and audience you wish to target.
• Interview the Client: Reach out to the company you want to showcase and ask relevant questions about their journey and achievements.
• Revision & Editing: Review your case study and ask for feedback. Include relevant quotes and CTAs to your case study.
• Publication & Distribution: Publish and share your case study on your website, social media channels and email list!
• Marketing & Repurposing: Turn your case study into a podcast, PDF, case study presentation and more. Share these materials with your sales and marketing team.

What Are the Advantages and Disadvantages of a Case Study?

Advantages of a case study:

• Case studies showcase a specific solution and outcome for specific customer challenges.
• It attracts potential customers with similar challenges.
• It builds trust and credibility with potential customers.
• It provides an in-depth analysis of your company’s problem-solving process.

Disadvantages of a case study:

• Limited applicability. Case studies are tailored to specific cases and may not apply to other businesses.
• It relies heavily on customer cooperation and willingness to share information.
• It stands a risk of becoming outdated as industries and customer needs evolve.

What Are the Types of Case Studies?

There are 7 main types of case studies. They include;

• Illustrative case study.
• Instrumental case study.
• Intrinsic case study.
• Descriptive case study.
• Explanatory case study.
• Exploratory case study.
• Collective case study.

How Long Should a Case Study Be?

The ideal length of your case study is between 500 - 1500 words or 1-3 pages. Certain factors like your target audience, goal or the amount of detail you want to share may influence the length of your case study. This infographic has powerful tips for designing winning case studies

What Is the Difference Between a Case Study and an Example?

Case studies provide a detailed narrative of how your product or service was used to solve a problem. Examples are general illustrations and are not necessarily real-life scenarios.

Case studies are often used for marketing purposes, attracting potential customers and building trust. Examples, on the other hand, are primarily used to simplify or clarify complex concepts.

Where Can I Find Case Study Examples?

You can easily find many case study examples online and in industry publications. Many companies, including Visme, share case studies on their websites to showcase how their products or services have helped clients achieve success. You can also search online libraries and professional organizations for case studies related to your specific industry or field.

If you need professionally-designed, customizable case study templates to create your own, Visme's template library is one of the best places to look. These templates include all the essential sections of a case study and high-quality content to help you create case studies that position your business as an industry leader.

Get More Out Of Your Case Studies With Visme

Case studies are an essential tool for converting potential customers into paying customers. By following the tips in this article, you can create compelling case studies that will help you build trust, establish credibility and drive sales.

Visme can help you create stunning case studies and other relevant marketing materials. With our easy-to-use platform, interactive features and analytics tools , you can increase your content creation game in no time.

There is no limit to what you can achieve with Visme. Connect with Sales to discover how Visme can boost your business goals.

Easily create beautiful case studies and more with Visme

Trusted by leading brands

Recommended content for you:

Create Stunning Content!

Design visual brand experiences for your business whether you are a seasoned designer or a total novice.

About the Author

Have a language expert improve your writing

Run a free plagiarism check in 10 minutes, generate accurate citations for free.

• Knowledge Base
• Research paper
• How to Write Recommendations in Research | Examples & Tips

How to Write Recommendations in Research | Examples & Tips

Published on September 15, 2022 by Tegan George . Revised on July 18, 2023.

Recommendations in research are a crucial component of your discussion section and the conclusion of your thesis , dissertation , or research paper .

As you conduct your research and analyze the data you collected , perhaps there are ideas or results that don’t quite fit the scope of your research topic. Or, maybe your results suggest that there are further implications of your results or the causal relationships between previously-studied variables than covered in extant research.

Instantly correct all language mistakes in your text

Upload your document to correct all your mistakes in minutes

Table of contents

What should recommendations look like, building your research recommendation, how should your recommendations be written, recommendation in research example, other interesting articles, frequently asked questions about recommendations.

Recommendations for future research should be:

• Concrete and specific
• Supported with a clear rationale
• Directly connected to your research

Overall, strive to highlight ways other researchers can reproduce or replicate your results to draw further conclusions, and suggest different directions that future research can take, if applicable.

Relatedly, when making these recommendations, avoid:

• Undermining your own work, but rather offer suggestions on how future studies can build upon it
• Suggesting recommendations actually needed to complete your argument, but rather ensure that your research stands alone on its own merits
• Using recommendations as a place for self-criticism, but rather as a natural extension point for your work

Don't submit your assignments before you do this

The academic proofreading tool has been trained on 1000s of academic texts. Making it the most accurate and reliable proofreading tool for students. Free citation check included.

Try for free

There are many different ways to frame recommendations, but the easiest is perhaps to follow the formula of research question   conclusion  recommendation. Here’s an example.

Conclusion An important condition for controlling many social skills is mastering language. If children have a better command of language, they can express themselves better and are better able to understand their peers. Opportunities to practice social skills are thus dependent on the development of language skills.

As a rule of thumb, try to limit yourself to only the most relevant future recommendations: ones that stem directly from your work. While you can have multiple recommendations for each research conclusion, it is also acceptable to have one recommendation that is connected to more than one conclusion.

These recommendations should be targeted at your audience, specifically toward peers or colleagues in your field that work on similar subjects to your paper or dissertation topic . They can flow directly from any limitations you found while conducting your work, offering concrete and actionable possibilities for how future research can build on anything that your own work was unable to address at the time of your writing.

See below for a full research recommendation example that you can use as a template to write your own.

Prevent plagiarism. Run a free check.

If you want to know more about AI for academic writing, AI tools, or research bias, make sure to check out some of our other articles with explanations and examples or go directly to our tools!

Research bias

• Survivorship bias
• Self-serving bias
• Availability heuristic
• Halo effect
• Hindsight bias
• Deep learning
• Generative AI
• Machine learning
• Reinforcement learning
• Supervised vs. unsupervised learning

 (AI) Tools

• Grammar Checker
• Paraphrasing Tool
• Text Summarizer
• AI Detector
• Plagiarism Checker
• Citation Generator

While it may be tempting to present new arguments or evidence in your thesis or disseration conclusion , especially if you have a particularly striking argument you’d like to finish your analysis with, you shouldn’t. Theses and dissertations follow a more formal structure than this.

All your findings and arguments should be presented in the body of the text (more specifically in the discussion section and results section .) The conclusion is meant to summarize and reflect on the evidence and arguments you have already presented, not introduce new ones.

The conclusion of your thesis or dissertation should include the following:

• A restatement of your research question
• A summary of your key arguments and/or results
• A short discussion of the implications of your research

For a stronger dissertation conclusion , avoid including:

• Important evidence or analysis that wasn’t mentioned in the discussion section and results section
• Generic concluding phrases (e.g. “In conclusion …”)
• Weak statements that undermine your argument (e.g., “There are good points on both sides of this issue.”)

Your conclusion should leave the reader with a strong, decisive impression of your work.

In a thesis or dissertation, the discussion is an in-depth exploration of the results, going into detail about the meaning of your findings and citing relevant sources to put them in context.

The conclusion is more shorter and more general: it concisely answers your main research question and makes recommendations based on your overall findings.

Cite this Scribbr article

If you want to cite this source, you can copy and paste the citation or click the “Cite this Scribbr article” button to automatically add the citation to our free Citation Generator.

George, T. (2023, July 18). How to Write Recommendations in Research | Examples & Tips. Scribbr. Retrieved September 27, 2024, from https://www.scribbr.com/dissertation/recommendations-in-research/

Is this article helpful?

Tegan George

Other students also liked, how to write a discussion section | tips & examples, how to write a thesis or dissertation conclusion, how to write a results section | tips & examples, what is your plagiarism score.

• All Categories
• Marketing Analytics Software

What Is a Case Study? How to Write, Examples, and Template

In this post

How to write a case study

Case study template, case study examples, types of case studies, what are the benefits of case studies , what are the limitations of case studies , case study vs. testimonial.

In today's marketplace, conveying your product's value through a compelling narrative is crucial to genuinely connecting with your customers.

Your business can use marketing analytics tools to understand what customers want to know about your product. Once you have this information, the next step is to showcase your product and its benefits to your target audience. This strategy involves a mix of data, analysis, and storytelling. Combining these elements allows you to create a narrative that engages your audience. So, how can you do this effectively?

What is a case study? 

A case study is a powerful tool for showcasing a business's success in helping clients achieve their goals. It's a form of storytelling that details real-world scenarios where a business implemented its solutions to deliver positive results for a client.

In this article, we explore the concept of a case study , including its writing process, benefits, various types, challenges, and more.

Understanding how to write a case study is an invaluable skill. You'll need to embrace decision-making – from deciding which customers to feature to designing the best format to make them as engaging as possible.  This can feel overwhelming in a hurry, so let's break it down.

Step 1: Reach out to the target persona

If you've been in business for a while, you have no shortage of happy customers. But w ith limited time and resources, you can't choose everyone.  So, take some time beforehand to flesh out your target buyer personas. 

Once you know precisely who you're targeting, go through your stable of happy customers to find a buyer representative of the audience you're trying to reach. The closer their problems, goals, and industries align, the more your case study will resonate.

What if you have more than one buyer persona? No problem. This is a common situation for companies because buyers comprise an entire committee. You might be marketing to procurement experts, executives, engineers, etc. Try to develop a case study tailored to each key persona. This might be a long-term goal, and that's fine. The better you can personalize the experience for each stakeholder, the easier it is to keep their attention.  

Here are a few considerations to think about before research:

• Products/services of yours the customer uses (and how familiar they are with them)
• The customer's brand recognition in the industry
• Whether the results they've achieved are specific and remarkable
• Whether they've switched from a competitor's product/service
• How closely aligned they are with your target audience

These items are just a jumping-off point as you develop your criteria.  Once you have a list, run each customer through it to determine your top targets. Approach the ones on the top (your "dream" case study subjects) and work your way down as needed.

Who to interview

You should consider interviewing top-level managers or executives because those are high-profile positions. But consider how close they are to your product and its results.

Focusing on an office manager or engineer who uses your product daily would be better. Look for someone with a courtside view of the effects.

The ways to request customer participation in case studies can vary, but certain principles can improve your chances:

• Make it easy for customers to work with you, respecting their valuable time. Be well-prepared and minimize their involvement.
• Emphasize how customers will benefit through increased publicity, revenue opportunities, or recognition for their success. 
• Acknowledge their contributions and showcase their achievements.
• Standardizing the request process with a script incorporating these principles can help your team consistently secure case study approvals and track performance.

Step 2: Prepare for the interview

Case study interviews are like school exams. The more prepared you are for them, the better they turn out. Preparing thoroughly also shows participants that you value their time. You don't waste precious minutes rehashing things you should have already known. You focus on getting the information you need as efficiently as possible.

You can conduct your case study interview in multiple formats, from exchanging emails to in-person interviews. This isn't a trivial decision.  As you'll see in the chart below, each format has its unique advantages and disadvantages. 

You'll also have to consider who will ask and answer the questions during your case study interview. It's wise to consider this while considering the case study format.  The number of participants factors into which format will work best. Pulling off an in-person interview becomes much harder if you're trying to juggle four or five people's busy schedules. Try a video conference instead.

Before interviewing your case study participant, it is crucial to identify the specific questions that need to be asked.  It's essential to thoroughly evaluate your collaboration with the client and understand how your product's contributions impact the company. 

Remember that structuring your case study is akin to crafting a compelling narrative. To achieve this, follow a structured approach:

• Beginning of your story. Delve into the customer's challenge that ultimately led them to do business with you. What were their problems like? What drove them to make a decision finally? Why did they choose you?
• The middle of the case study.  Your audience also wants to know about the experience of working with you. Your customer has taken action to address their problems. What happened once you got on board?
• An ending that makes you the hero.  Describe the specific results your company produced for the customer. How has the customer's business (and life) changed once they implemented your solution?

Sample questions for the case study interview

If you're preparing for a case study interview, here are some sample case study research questions to help you get started:

• What challenges led you to seek a solution?
• When did you realize the need for immediate action? Was there a tipping point?
• How did you decide on the criteria for choosing a B2B solution, and who was involved?
• What set our product or service apart from others you considered?
• How was your experience working with us post-purchase?
• Were there any pleasant surprises or exceeded expectations during our collaboration?
• How smoothly did your team integrate our solution into their workflows?
• How long before you started seeing positive results?
• How have you benefited from our products or services?
• How do you measure the value our product or service provides?

Step 3: Conduct the interview

Preparing for case study interviews can be different from everyday conversations. Here are some tips to keep in mind:

• Create a comfortable atmosphere.  Before diving into the discussion, talk about their business and personal interests. Ensure everyone is at ease, and address any questions or concerns.
• Prioritize key questions.  Lead with your most crucial questions to respect your customer's time. Interview lengths can vary, so starting with the essentials ensures you get the vital information.
• Be flexible.  Case study interviews don't have to be rigid. If your interviewee goes "off script," embrace it. Their spontaneous responses often provide valuable insights.
• Record the interview.  If not conducted via email, ask for permission to record the interview. This lets you focus on the conversation and capture valuable quotes without distractions.

Step 4: Figure out who will create the case study

When creating written case studies for your business, deciding who should handle the writing depends on cost, perspective, and revisions.

Outsourcing might be pricier, but it ensures a professionally crafted outcome. On the other hand, in-house writing has its considerations, including understanding your customers and products. 

Technical expertise and equipment are needed for video case studies, which often leads companies to consider outsourcing due to production and editing costs. 

Tip: When outsourcing work, it's essential to clearly understand pricing details to avoid surprises and unexpected charges during payment.

Step 5: Utilize storytelling

Understanding and applying storytelling elements can make your case studies unforgettable, offering a competitive edge. 

Source: The Framework Bank

Every great study follows a narrative arc (also called a "story arc"). This arc represents how a character faces challenges, struggles against raising stakes, and encounters a formidable obstacle before the tension resolves.

In a case study narrative, consider:

• Exposition. Provide background information about the company, revealing their "old life" before becoming your customer.
• Inciting incident. Highlight the problem that drove the customer to seek a solution, creating a sense of urgency.
• Obstacles (rising action). Describe the customer's journey in researching and evaluating solutions, building tension as they explore options.
• Midpoint. Explain what made the business choose your product or service and what set you apart.
• Climax. Showcase the success achieved with your product.
• Denouement. Describe the customer's transformed business and end with a call-to-action for the reader to take the next step.

Step 6: Design the case study

The adage "Don't judge a book by its cover" is familiar, but people tend to do just that quite often!

A poor layout can deter readers even if you have an outstanding case study. To create an engaging case study, follow these steps:

• Craft a compelling title. Just like you wouldn't read a newspaper article without an eye-catching headline, the same goes for case studies. Start with a title that grabs attention.
• Organize your content. Break down your content into different sections, such as challenges, results, etc. Each section can also include subsections. This case study approach divides the content into manageable portions, preventing readers from feeling overwhelmed by lengthy blocks of text.
• Conciseness is key. Keep your case study as concise as possible. The most compelling case studies are precisely long enough to introduce the customer's challenge, experience with your solution, and outstanding results. Prioritize clarity and omit any sections that may detract from the main storyline.
• Utilize visual elements. To break up text and maintain reader interest, incorporate visual elements like callout boxes, bulleted lists, and sidebars.
• Include charts and images. Summarize results and simplify complex topics by including pictures and charts. Visual aids enhance the overall appeal of your case study.
• Embrace white space. Avoid overwhelming walls of text to prevent reader fatigue. Opt for plenty of white space, use shorter paragraphs, and employ subsections to ensure easy readability and navigation.
• Enhance video case studies. In video case studies, elements like music, fonts, and color grading are pivotal in setting the right tone. Choose music that complements your message and use it strategically throughout your story. Carefully select fonts to convey the desired style, and consider how lighting and color grading can influence the mood. These elements collectively help create the desired tone for your video case study.

Step 7: Edits and revisions

Once you've finished the interview and created your case study, the hardest part is over. Now's the time for editing and revision. This might feel frustrating for impatient B2B marketers, but it can turn good stories into great ones.

Ideally, you'll want to submit your case study through two different rounds of editing and revisions:

• Internal review. Seek feedback from various team members to ensure your case study is captivating and error-free. Gather perspectives from marketing, sales, and those in close contact with customers for well-rounded insights. Use patterns from this feedback to guide revisions and apply lessons to future case studies.
• Customer feedback. Share the case study with customers to make them feel valued and ensure accuracy. Let them review quotes and data points, as they are the "heroes" of the story, and their logos will be prominently featured. This step maintains positive customer relationships.

Case study mistakes to avoid

• Ensure easy access to case studies on your website.
• Spotlight the customer, not just your business.
• Tailor each case study to a specific audience.
• Avoid excessive industry jargon in your content.

Step 8: Publishing

Take a moment to proofread your case study one more time carefully. Even if you're reasonably confident you've caught all the errors, it's always a good idea to check. Your case study will be a valuable marketing tool for years, so it's worth the investment to ensure it's flawless. Once done, your case study is all set to go!

Consider sharing a copy of the completed case study with your customer as a thoughtful gesture. They'll likely appreciate it; some may want to keep it for their records. After all, your case study wouldn't have been possible without their help, and they deserve to see the final product.

Where you publish your case study depends on its role in your overall marketing strategy. If you want to reach as many people as possible with your case study, consider publishing it on your website and social media platforms. 

Tip: Some companies prefer to keep their case studies exclusive, making them available only to those who request them. This approach is often taken to control access to valuable information and to engage more deeply with potential customers who express specific interests. It can create a sense of exclusivity and encourage interested parties to engage directly with the company.

Step 9: Case study distribution

When sharing individual case studies, concentrate on reaching the audience with the most influence on purchasing decisions

Here are some common distribution channels to consider:

• Sales teams. Share case studies to enhance customer interactions, retention , and upselling among your sales and customer success teams. Keep them updated on new studies and offer easily accessible formats like PDFs or landing page links.
• Company website. Feature case studies on your website to establish authority and provide valuable information to potential buyers. Organize them by categories such as location, size, industry, challenges, and products or services used for effective presentation.
• Events. Use live events like conferences and webinars to distribute printed case study copies, showcase video case studies at trade show booths, and conclude webinars with links to your case study library. This creative approach blends personal interactions with compelling content.
• Industry journalists. Engage relevant industry journalists to gain media coverage by identifying suitable publications and journalists covering related topics. Building relationships is vital, and platforms like HARO (Help A Reporter Out) can facilitate connections, especially if your competitors have received coverage before.

Want to learn more about Marketing Analytics Software? Explore Marketing Analytics products.

It can seem daunting to transform the information you've gathered into a cohesive narrative.  We’ve created a versatile case study template that can serve as a solid starting point for your case study.

With this template, your business can explore any solutions offered to satisfied customers, covering their background, the factors that led them to choose your services, and their outcomes.

The template boasts a straightforward design, featuring distinct sections that guide you in effectively narrating your and your customer's story. However, remember that limitless ways to showcase your business's accomplishments exist.

To assist you in this process, here's a breakdown of the recommended sections to include in a case study:

• Title.  Keep it concise. Create a brief yet engaging project title summarizing your work with your subject. Consider your title like a newspaper headline; do it well, and readers will want to learn more. 
• Subtitle . Use this section to elaborate on the achievement briefly. Make it creative and catchy to engage your audience.
• Executive summary.  Use this as an overview of the story, followed by 2-3 bullet points highlighting key success metrics.
• Challenges and objectives. This section describes the customer's challenges before adopting your product or service, along with the goals or objectives they sought to achieve.
• How product/service helped.  A paragraph explaining how your product or service addressed their problem.
• Testimonials.  Incorporate short quotes or statements from the individuals involved in the case study, sharing their perspectives and experiences.
• Supporting visuals.  Include one or two impactful visuals, such as graphs, infographics, or highlighted metrics, that reinforce the narrative.
• Call to action (CTA).  If you do your job well, your audience will read (or watch) your case studies from beginning to end. They are interested in everything you've said. Now, what's the next step they should take to continue their relationship with you? Give people a simple action they can complete. 

Case studies are proven marketing strategies in a wide variety of B2B industries. Here are just a few examples of a case study:

• Amazon Web Services, Inc.  provides companies with cloud computing platforms and APIs on a metered, pay-as-you-go basis. This case study example illustrates the benefits Thomson Reuters experienced using AWS.
• LinkedIn Marketing Solutions combines captivating visuals with measurable results in the case study created for BlackRock. This case study illustrates how LinkedIn has contributed to the growth of BlackRock's brand awareness over the years. 
• Salesforce , a sales and marketing automation SaaS solutions provider, seamlessly integrates written and visual elements to convey its success stories with Pepe Jeans. This case study effectively demonstrates how Pepe Jeans is captivating online shoppers with immersive and context-driven e-commerce experiences through Salesforce.
• HubSpot offers a combination of sales and marketing tools. Their case study demonstrates the effectiveness of its all-in-one solutions. These typically focus on a particular client's journey and how HubSpot helped them achieve significant results.

There are two different types of case studies that businesses might utilize:

Written case studies 

Written case studies offer readers a clear visual representation of data, which helps them quickly identify and focus on the information that matters most. 

Printed versions of case studies find their place at events like trade shows, where they serve as valuable sales collateral to engage prospective clients.  Even in the digital age, many businesses provide case studies in PDF format or as web-based landing pages, improving accessibility for their audience. 

Note: Landing pages , in particular, offer the flexibility to incorporate rich multimedia content, including images, charts, and videos. This flexibility in design makes landing pages an attractive choice for presenting detailed content to the audience.

Written case study advantages

Here are several significant advantages to leveraging case studies for your company:

• Hyperlink accessibility.  Whether in PDF or landing page format, written case studies allow for embedded hyperlinks, offering prospects easy access to additional information and contact forms.
• Flexible engagement.  Unlike video case studies, which may demand in-person arrangements, written case studies can be conducted via phone or video streaming, reducing customer commitment and simplifying scheduling.
• Efficient scanning . Well-structured written case studies with a scannable format cater to time-strapped professionals. Charts and callout boxes with key statistics enhance the ease of information retrieval.
• Printable for offline use.  Written case studies can be effortlessly printed and distributed at trade shows, sales meetings, and live events. This tangible format accommodates those who prefer physical materials and provides versatility in outreach, unlike video content, which is less portable.

Written case study disadvantages

Here are some drawbacks associated with the use of case studies:

• Reduced emotional impact.  Written content lacks the emotional punch of live video testimonials, which engage more senses and emotions, making a stronger connection.
• Consider time investment.  Creating a compelling case study involves editing, proofreading, and design collaboration, with multiple revisions commonly required before publication.
• Challenges in maintaining attention.  Attention spans are short in today's ad-saturated world. Using graphics, infographics, and videos more often is more powerful to incite the right emotions in customers.

Video case studies

Video case studies are the latest marketing trend. Unlike in the past, when video production was costly, today's tools make it more accessible for users to create and edit their videos. However, specific technical requirements still apply.

Like written case studies, video case studies delve into a specific customer's challenges and how your business provides solutions. Yet, the video offers a more profound connection by showcasing the person who faced and conquered the problem.

Video case studies can boost brand exposure when shared on platforms like YouTube. For example, Slack's engaging case study video with Sandwich Video illustrates how Slack transformed its workflow and adds humor, which can be challenging in written case studies focused on factual evidence.

Source : YouTube

This video case study has garnered nearly a million views on YouTube.

Video case study advantages

Here are some of the top advantages of video case studies. While video testimonials take more time, the payoff can be worth it. 

• Humanization and authenticity.  Video case studies connect viewers with real people, adding authenticity and fostering a stronger emotional connection.
• Engaging multiple senses.  They engage both auditory and visual senses, enhancing credibility and emotional impact. Charts, statistics, and images can also be incorporated.
• Broad distribution.  Videos can be shared on websites, YouTube, social media, and more, reaching diverse audiences and boosting engagement, especially on social platforms.

Video case study disadvantages

Before fully committing to video testimonials, consider the following:

• Technical expertise and equipment.  Video production requires technical know-how and equipment, which can be costly. Skilled video editing is essential to maintain a professional image. While technology advances, producing amateurish videos may harm your brand's perception.
• Viewer convenience.  Some prospects prefer written formats due to faster reading and ease of navigation. Video typically requires sound, which can be inconvenient for viewers in specific settings. Many people may not have headphones readily available to watch your content.
• Demand on case study participants.  On-camera interviews can be time-consuming and location-dependent, making scheduling challenging for case study participants. Additionally, being on screen for a global audience may create insecurities and performance pressure.
• Comfort on camera.  Not everyone feels at ease on camera. Nervousness or a different on-screen persona can impact the effectiveness of the testimonial, and discovering this late in the process can be problematic.

Written or video case studies: Which is right for you?

Now that you know the pros and cons of each, how do you choose which is right for you?

One of the most significant factors in doing video case studies can be the technical expertise and equipment required for a high level of production quality. Whether you have the budget to do this in-house or hire a production company can be one of the major deciding factors.

Still, written or video doesn't have to be an either-or decision. Some B2B companies are using both formats. They can complement each other nicely, minimizing the downsides mentioned above and reaching your potential customers where they prefer.

Let's say you're selling IT network security. What you offer is invaluable but complicated. You could create a short (three- or four-minute) video case study to get attention and touch on the significant benefits of your services. This whets the viewer's appetite for more information, which they could find in a written case study that supplements the video.

Should you decide to test the water in video case studies, test their effectiveness among your target audience. See how well they work for your company and sales team. And, just like a written case study, you can always find ways to improve your process as you continue exploring video case studies.

Case studies offer several distinctive advantages, making them an ideal tool for businesses to market their products to customers. However, their benefits extend beyond these qualities. 

Here's an overview of all the advantages of case studies:

Valuable sales support

Case studies serve as a valuable resource for your sales endeavors. Buyers frequently require additional information before finalizing a purchase decision. These studies provide concrete evidence of your product or service's effectiveness, assisting your sales representatives in closing deals more efficiently, especially with customers with lingering uncertainties.

Validating your value

Case studies serve as evidence of your product or service's worth or value proposition , playing a role in building trust with potential customers. By showcasing successful partnerships, you make it easier for prospects to place trust in your offerings. This effect is particularly notable when the featured customer holds a reputable status.

Unique and engaging content

By working closely with your customer success teams, you can uncover various customer stories that resonate with different prospects. Case studies allow marketers to shape product features and benefits into compelling narratives. 

Each case study's distinctiveness, mirroring the uniqueness of every customer's journey, makes them a valuable source of relatable and engaging content. Storytelling possesses the unique ability to connect with audiences on an emotional level, a dimension that statistics alone often cannot achieve. 

Spotlighting valuable customers

Case studies provide a valuable platform for showcasing your esteemed customers. Featuring them in these studies offers a chance to give them visibility and express your gratitude for the partnership, which can enhance customer loyalty . Depending on the company you are writing about, it can also demonstrate the caliber of your business.

Now is the time to get SaaS-y news and entertainment with our 5-minute newsletter,   G2 Tea , featuring inspiring leaders, hot takes, and bold predictions. Subscribe below!

It's important to consider limitations when designing and interpreting the results of case studies. Here's an overview of the limitations of case studies:

Challenges in replication

Case studies often focus on specific individuals, organizations, or situations, making generalizing their findings to broader populations or contexts challenging. 

Time-intensive process

Case studies require a significant time investment. The extensive data collection process and the need for comprehensive analysis can be demanding, especially for researchers who are new to this method.

Potential for errors

Case studies can be influenced by memory and judgment, potentially leading to inaccuracies. Depending on human memory to reconstruct a case's history may result in variations and potential inconsistencies in how individuals recall past events. Additionally, bias may emerge, as individuals tend to prioritize what they consider most significant, which could limit their consideration of alternative perspectives.

Challenges in verification

Confirming results through additional research can present difficulties. This complexity arises from the need for detailed and extensive data in the initial creation of a case study. Consequently, this process requires significant effort and a substantial amount of time.

While looking at case studies, you may have noticed a quote. This type of quote is considered a testimonial, a key element of case studies.

If a customer's quote proves that your brand does what it says it will or performs as expected, you may wonder: 'Aren't customer testimonials and case studies the same thing?' Not exactly.

Testimonials are brief endorsements designed to establish trust on a broad scale. In contrast, case studies are detailed narratives that offer a comprehensive understanding of how a product or service addresses a specific problem, targeting a more focused audience. 

Crafting case studies requires more resources and a structured approach than testimonials. Your selection between the two depends on your marketing objectives and the complexity of your product or service.

Case in point!

Case studies are among a company's most effective tools. You're  well on your way to mastering them.

Today's buyers are tackling much of the case study research methodology independently. Many are understandably skeptical before making a buying decision. By connecting them with multiple case studies, you can prove you've gotten the results you say you can. There's hardly a better way to boost your credibility and persuade them to consider your solution.

Case study formats and distribution methods might change as technology evolves. However, the fundamentals that make them effective—knowing how to choose subjects, conduct interviews, and structure everything to get attention—will serve you for as long as you're in business. 

We covered a ton of concepts and resources, so go ahead and bookmark this page. You can refer to it whenever you have questions or need a refresher.

Dive into market research to uncover customer preferences and spending habits.

Kristen McCabe

Kristen’s is a former senior content marketing specialist at G2. Her global marketing experience extends from Australia to Chicago, with expertise in B2B and B2C industries. Specializing in content, conversions, and events, Kristen spends her time outside of work time acting, learning nature photography, and joining in the #instadog fun with her Pug/Jack Russell, Bella. (she/her/hers)

Explore More G2 Articles

Notebook is a mobile first digital learning product that combines videos and text content to deliver best-in-class education according to Board curricula.

SkillsPro Essential

SkillsPro Essential is an innovative and comprehensive tech product that has been meticulously designed to provide cutting-edge upskilling courses for individuals seeking to embark on a journey of self-improvement. 

SkillsPro Tech

SkillsPro Tech, the ultimate destination for empowering your tech-driven future! Our cutting-edge tech product is crafted to offer comprehensive upskilling courses.

SkillsBasic

SkillsBasic  the magical world of online learning designed to nurture and ignite young minds aged 4 to 12 years. We are on a mission to provide kids with the fundamental skills they need to unlock their full potential and pave the way for a bright future.

Data Analytics Program

Executive program in data analytics & data science, executive program in full stack .net software development.

Have you ever bought something — within the last 10 years or so — without reading its reviews or without a recommendation or prior experience of using it?

If the answer is no — or at least, rarely — you get my point.

Positive reviews matter for selling to regular customers, and for B2B or SaaS businesses, detailed case studies are important too.

Wondering how to craft a compelling case study ? No worries—I’ve got you covered with 15 marketing case study templates , helpful tips, and examples to ensure your case study converts effectively.

Click to jump ahead:

What is a case study?

What to include in a professional case study, business case study examples, simple case study examples, marketing case study examples, sales case study examples.

• Case study FAQs

A case study is an in-depth, detailed analysis of a specific real-world situation. For example, a case study can be about an individual, group, event, organization, or phenomenon. The purpose of a case study is to understand its complexities and gain insights into a particular instance or situation.

In the context of a business, however, case studies take customer success stories and explore how they use your product to help them achieve their business goals.

As well as being valuable marketing tools , case studies are a good way to evaluate your product as it allows you to objectively examine how others are using it.

It’s also a good way to interview your customers about why they work with you.

Related: What is a Case Study? [+6 Types of Case Studies]

A professional case study showcases how your product or services helped potential clients achieve their business goals. You can also create case studies of internal, successful marketing projects. A professional case study typically includes:

• Company background and history
• The challenge
• How you helped
• Specific actions taken
• Visuals or Data
• Client testimonials

Here’s an example of a case study template:

Whether you’re a B2B or B2C company, business case studies can be a powerful resource to help with your sales, marketing, and even internal departmental awareness.

Business and business management case studies should encompass strategic insights alongside anecdotal and qualitative findings, like in the business case study examples below.

Conduct a B2B case study by researching the company holistically

When it comes to writing a case study, make sure you approach the company holistically and analyze everything from their social media to their sales.

Think about every avenue your product or service has been of use to your case study company, and ask them about the impact this has had on their wider company goals.

In business case study examples like the one above, we can see that the company has been thought about holistically simply by the use of icons.

By combining social media icons with icons that show in-person communication we know that this is a well-researched and thorough case study.

This case study report example could also be used within an annual or end-of-year report.

Highlight the key takeaway from your marketing case study

To create a compelling case study, identify the key takeaways from your research. Use catchy language to sum up this information in a sentence, and present this sentence at the top of your page.

This is “at a glance” information and it allows people to gain a top-level understanding of the content immediately. 

You can use a large, bold, contrasting font to help this information stand out from the page and provide interest.

Learn  how to choose fonts  effectively with our Venngage guide and once you’ve done that.

Upload your fonts and  brand colors  to Venngage using the  My Brand Kit  tool and see them automatically applied to your designs.

The heading is the ideal place to put the most impactful information, as this is the first thing that people will read.

In this example, the stat of “Increase[d] lead quality by 90%” is used as the header. It makes customers want to read more to find out how exactly lead quality was increased by such a massive amount.

If you’re conducting an in-person interview, you could highlight a direct quote or insight provided by your interview subject.

Pick out a catchy sentence or phrase, or the key piece of information your interview subject provided and use that as a way to draw a potential customer in.

Use charts to visualize data in your business case studies

Charts are an excellent way to visualize data and to bring statistics and information to life. Charts make information easier to understand and to illustrate trends or patterns.

Making charts is even easier with Venngage.

In this consulting case study example, we can see that a chart has been used to demonstrate the difference in lead value within the Lead Elves case study.

Adding a chart here helps break up the information and add visual value to the case study. 

Using charts in your case study can also be useful if you’re creating a project management case study.

You could use a Gantt chart or a project timeline to show how you have managed the project successfully.

Use direct quotes to build trust in your marketing case study

To add an extra layer of authenticity you can include a direct quote from your customer within your case study.

According to research from Nielsen , 92% of people will trust a recommendation from a peer and 70% trust recommendations even if they’re from somebody they don’t know.

So if you have a customer or client who can’t stop singing your praises, make sure you get a direct quote from them and include it in your case study.

You can either lift part of the conversation or interview, or you can specifically request a quote. Make sure to ask for permission before using the quote.

This design uses a bright contrasting speech bubble to show that it includes a direct quote, and helps the quote stand out from the rest of the text.

This will help draw the customer’s attention directly to the quote, in turn influencing them to use your product or service.

Less is often more, and this is especially true when it comes to creating designs. Whilst you want to create a professional-looking, well-written and design case study – there’s no need to overcomplicate things.

These simple case study examples show that smart clean designs and informative content can be an effective way to showcase your successes.

Use colors and fonts to create a professional-looking case study

Business case studies shouldn’t be boring. In fact, they should be beautifully and professionally designed.

This means the normal rules of design apply. Use fonts, colors, and icons to create an interesting and visually appealing case study.

In this case study example, we can see how multiple fonts have been used to help differentiate between the headers and content, as well as complementary colors and eye-catching icons.

Marketing case studies are incredibly useful for showing your marketing successes. Every successful marketing campaign relies on influencing a consumer’s behavior, and a great case study can be a great way to spotlight your biggest wins.

In the marketing case study examples below, a variety of designs and techniques to create impactful and effective case studies.

Show off impressive results with a bold marketing case study

Case studies are meant to show off your successes, so make sure you feature your positive results prominently. Using bold and bright colors as well as contrasting shapes, large bold fonts, and simple icons is a great way to highlight your wins.

In well-written case study examples like the one below, the big wins are highlighted on the second page with a bright orange color and are highlighted in circles.

Making the important data stand out is especially important when attracting a prospective customer with marketing case studies.

Use a simple but clear layout in your case study

Using a simple layout in your case study can be incredibly effective, like in the example of a case study below.

Keeping a clean white background, and using slim lines to help separate the sections is an easy way to format your case study.

Making the information clear helps draw attention to the important results, and it helps improve the  accessibility of the design .

Business case study examples like this would sit nicely within a larger report, with a consistent layout throughout.

Use visuals and icons to create an engaging and branded business case study

Nobody wants to read pages and pages of text — and that’s why Venngage wants to help you communicate your ideas visually.

Using icons, graphics, photos, or patterns helps create a much more engaging design. 

With this Blue Cap case study icons, colors, and impactful pattern designs have been used to create an engaging design that catches your eye.

Use a monochromatic color palette to create a professional and clean case study

Let your research shine by using a monochromatic and minimalistic color palette.

By sticking to one color, and leaving lots of blank space you can ensure your design doesn’t distract a potential customer from your case study content.

In this case study on Polygon Media, the design is simple and professional, and the layout allows the prospective customer to follow the flow of information.

The gradient effect on the left-hand column helps break up the white background and adds an interesting visual effect.

Did you know you can generate an accessible color palette with Venngage? Try our free accessible color palette generator today and create a case study that delivers and looks pleasant to the eye:

Add long term goals in your case study

When creating a case study it’s a great idea to look at both the short term and the long term goals of the company to gain the best understanding possible of the insights they provide.

Short-term goals will be what the company or person hopes to achieve in the next few months, and long-term goals are what the company hopes to achieve in the next few years.

Check out this modern pattern design example of a case study below:

In this case study example, the short and long-term goals are clearly distinguished by light blue boxes and placed side by side so that they are easy to compare.

Use a strong introductory paragraph to outline the overall strategy and goals before outlining the specific short-term and long-term goals to help with clarity.

This strategy can also be handy when creating a consulting case study.

Use data to make concrete points about your sales and successes

When conducting any sort of research stats, facts, and figures are like gold dust (aka, really valuable).

Being able to quantify your findings is important to help understand the information fully. Saying sales increased 10% is much more effective than saying sales increased.

While sales dashboards generally tend it make it all about the numbers and charts, in sales case study examples, like this one, the key data and findings can be presented with icons. This contributes to the potential customer’s better understanding of the report.

They can clearly comprehend the information and it shows that the case study has been well researched.

Use emotive, persuasive, or action based language in your marketing case study

Create a compelling case study by using emotive, persuasive and action-based language when customizing your case study template.

In this well-written case study example, we can see that phrases such as “Results that Speak Volumes” and “Drive Sales” have been used.

Using persuasive language like you would in a blog post. It helps inspire potential customers to take action now.

Keep your potential customers in mind when creating a customer case study for marketing

82% of marketers use case studies in their marketing  because it’s such an effective tool to help quickly gain customers’ trust and to showcase the potential of your product.

Why are case studies such an important tool in content marketing?

By writing a case study you’re telling potential customers that they can trust you because you’re showing them that other people do.

Not only that, but if you have a SaaS product, business case studies are a great way to show how other people are effectively using your product in their company.

In this case study, Network is demonstrating how their product has been used by Vortex Co. with great success; instantly showing other potential customers that their tool works and is worth using.

Related: 10+ Case Study Infographic Templates That Convert

Case studies are particularly effective as a sales technique.

A sales case study is like an extended customer testimonial, not only sharing opinions of your product – but showcasing the results you helped your customer achieve.

Make impactful statistics pop in your sales case study

Writing a case study doesn’t mean using text as the only medium for sharing results.

You should use icons to highlight areas of your research that are particularly interesting or relevant, like in this example of a case study:

Icons are a great way to help summarize information quickly and can act as visual cues to help draw the customer’s attention to certain areas of the page.

In some of the business case study examples above, icons are used to represent the impressive areas of growth and are presented in a way that grabs your attention.

Use high contrast shapes and colors to draw attention to key information in your sales case study

Help the key information stand out within your case study by using high contrast shapes and colors.

Use a complementary or contrasting color, or use a shape such as a rectangle or a circle for maximum impact.

This design has used dark blue rectangles to help separate the information and make it easier to read.

Coupled with icons and strong statistics, this information stands out on the page and is easily digestible and retainable for a potential customer.

Case study examples summary

Once you have created your case study, it’s best practice to update your examples on a regular basis to include up-to-date statistics, data, and information.

You should update your business case study examples often if you are sharing them on your website .

It’s also important that your case study sits within your brand guidelines – find out how Venngage’s My Brand Kit tool can help you create consistently branded case study templates.

Case studies are important marketing tools – but they shouldn’t be the only tool in your toolbox. Content marketing is also a valuable way to earn consumer trust.

Case study FAQ s

Why should you write a case study.

Case studies are an effective marketing technique to engage potential customers and help build trust.

By producing case studies featuring your current clients or customers, you are showcasing how your tool or product can be used. You’re also showing that other people endorse your product.

In addition to being a good way to gather positive testimonials from existing customers, business case studies are good educational resources and can be shared amongst your company or team, and used as a reference for future projects.

How should you write a case study?

To create a great case study, you should think strategically. The first step, before starting your case study research, is to think about what you aim to learn or what you aim to prove.

You might be aiming to learn how a company makes sales or develops a new product. If this is the case, base your questions around this.

You can learn more about writing a case study  from our extensive guide.

Related: How to Present a Case Study like a Pro (With Examples)

Some good questions you could ask would be:

• Why do you use our tool or service?
• How often do you use our tool or service?
• What does the process of using our product look like to you?
• If our product didn’t exist, what would you be doing instead?
• What is the number one benefit you’ve found from using our tool?

You might also enjoy:

• 12 Essential Consulting Templates For Marketing, Planning and Branding
• Best Marketing Strategies for Consultants and Freelancers in 2019 [Study + Infographic]

Discover popular designs

Infographic maker

Brochure maker

White paper online

Newsletter creator

Flyer maker

Timeline maker

Letterhead maker

Mind map maker

Ebook maker

3 Tips To Create Case Studies That Turn Prospects Into Buyers [Examples]

• by Ann Gynn
• | Published: June 14, 2023
• | Content Creation

Case studies prevent your prospects from getting stuck in the funnel.

That’s probably why almost two-thirds of B2B content marketers use the tactic.

Case studies tell the story the prospective customer wants to know. Do you understand their pain points or need? Do you have a solution? Does it really deliver results? And case studies give those answers in a way that communicates credibility that an intrusive ad designed to sell, sell, sell could never do.

However, case studies can be a challenge to produce. You must talk to the sales and product teams to know which clients have had the best success stories. Then, you must track down the right person at the client company to get permission to tell their story. Plus, you almost always need numbers to prove the case, and they often aren’t easy to find.

But it’s worth the effort. In the 2023 B2B Content Marketing Benchmark, Budgets, and Trends , 36% of marketers cite case studies as effective – that’s more than long articles, podcasts and other audio content, infographics, livestreaming content, and print magazines and books. (And the number who cited case studies as effective was the same as those who said video – a rapidly growing tactic – was effective.)

Let these three strategies and four examples help you develop case studies that will shake your prospects from the middle of the funnel and turn them into customers.

1. Tell a story where your client – not your brand – stars

Content marketing should always focus on the audience, not the brand. In case studies, customer-centered storytelling is even more important. After all, prospects consume case studies because they want to see what happened with people/companies in similar industries and circumstances engaged with your brand.

If the case study sounds too promotional, it will lead the reader to think it’s just another ad where the company positions itself to sound great or, even worse – too good to be true.

By making your client the hero or star of the story, you make the necessary shift in the narrative and indicate to the reader that your brand is all about the clients and customers.

Xerox makes itself a secondary character in both of these case studies. In this PDF compilation of case studies for banking and financial markets , the cover includes the Xerox logo lower right. It also does a nice job of using a high-impact, non-product-related image. Unfortunately, it opted for a boring label-type headline.

On the next page, the case study collection includes a high-level overview and doesn’t mention Xerox or include the logo. The language emphasizes the client’s importance by using “you” throughout the text and opting for first person (our) in only a single reference.

“Customer loyalty is heavily influenced by how effectively you communicate, and that, in turn, depends on how easily you can make information flow …”

After this introduction, it shares eight case studies in a simple format that highlights points readers most likely want to know – the challenge, solution, and results. Xerox explains the solutions from the perspective of the client with no mention of its product names. For example, among the bullets in the solution category: “A single provider with dedicated operators to centrally manage the digitalisation (scanning and indexing) and onward distribution of incoming mail.”

TIP: Xerox doesn’t even name the “stars” of this case study collection. Many companies can’t identify their clients by name, and these case studies illustrate how to convey the need-to-know information to the readers without disclosing who the customer is.

While Xerox did a nice job focusing on the client, not the vendor, it didn’t tell engaging stories. But it did in the video version of this case study – Buena Park School District: Modernizing communication and today’s classroom with Xerox® Workflow Central .

During the 2.5-minute video, the district’s COO (someone likely involved in the decision-making and who uses the product) and a secretary (a frontline user) tell the story of their school system, its multi-language demographics, and the importance of communication. They also go into detail about how Xerox’s software product enables them to communicate with parents in multiple languages more effectively and efficiently.

The B-roll video includes screenshots of the Xerox product in use, but it also showcases the district, teachers, students, and more.

By having the school district employees tell their story and their experience, Xerox crafts a more interesting story and fosters a more credible case study. (Mysteriously, it chose not to use those interviews in the text version of the case study.

2. Create a familiar structure

Case studies aren’t the place for innovative story structures. All readers consume case studies to find the answers to the same questions: Do you know and understand their problem? Do you have a solution for it? Does that solution work?

They also may be doing comparison shopping through case studies. If readers have to work harder to figure out what they need to know from a creative story structure, they’re likely to move on to easier-to-read case studies.

Cognizant, a tech business, hosts a plethora of case studies across the industries it serves. It follows a familiar structure – challenge, approach, and business outcomes – for each.

In this example , they tell the story of their client, Planned Administrators, Inc., operating a legacy customer service application that didn’t integrate with its core administrative services. The study explains how that was problematic and details the customer’s goals – “reduce time and costs while improving operational efficiency to deliver a modern customer experience and support new lines of businesses.”

Then, it explains how Cognizant provided the solution. But the explanation didn’t stop with the solution’s description; the company also details how it helped Planned Administrators implement the new technology . That’s an important detail for anyone considering Cognizant’s solutions.

Though they don’t label it “business outcomes” in the web version of the case study, they clearly detail the impact in the descriptive concluding text and break out the all-important numbers – 10% reduction in call time, $80,000 saved in annual IT maintenance costs, $138,500 in annual fees eliminated.

Cognizant links to a four-page PDF of the case study for visitors interested in learning the full story. It includes much of the same information but adds quotes from the client and a sidebar box (shown below) that summarizes key details – industry, location, challenges, products and services, and success highlights. That’s especially helpful for skimming readers.

3. Make it visually engaging

Just because the story structure should be standard doesn’t mean the format must be. Readers usually don’t want to read paragraph after paragraph. It doesn’t stimulate their brain. And some readers might learn better through visuals.

So think about how to engage the audience – from videos and images to text design (headers, callout boxes, etc.). Create a design that allows the at-a-glance readers to jump around and get the highlights and gives the whole story to readers who want to consume every bit of information.

Superior Essex, a manufacturer of cables, opted for an interactive case study experience to tell the story of its work for the corporate space of Delos, a wellness real estate firm. (I encourage you to click through the full interactive case study, as describing it can’t adequately capture the experience.)

Superior Essex tells a similar story – explaining the client, its needs/pain points, solutions, results, and a testimonial quote. But it does it in a visually engaging way where readers can opt to move the story along on their own.

For example, this screen allows the reader to click through each solution with the product names identified.

On this page, the reader can learn about the impact on the company (as shown below), then simply click on the tabs (or advance the arrow at the bottom) to learn about the impact on the occupants and the environment.

TIP: Don’t forget the readers who might need to download the PDF to share with their buying teams or executives. Superior Essex condensed the interactive content into a simpler, more traditional three-page PDF .

Create convincing case studies that motivate buyers

Case studies help content marketers achieve some of their most common goals – building/growing credibility/trust, educating audiences, and generating and nurturing leads. But the power of case studies can be maximized if you make your clients the star, give your readers what they want, and throw in a dash of surprise and visual interest.

Cover image by Joseph Kalinowski/Content Marketing Institute

How To Make Recommendation in Case Study (With Examples)

After analyzing your case study’s problem and suggesting possible courses of action , you’re now ready to conclude it on a high note. 

But first, you need to write your recommendation to address the problem. In this article, we will guide you on how to make a recommendation in a case study. 

Table of Contents

What is recommendation in case study, what is the purpose of recommendation in the case study, 1. review your case study’s problem, 2. assess your case study’s alternative courses of action, 3. pick your case study’s best alternative course of action, 4. explain in detail why you recommend your preferred course of action, examples of recommendations in case study, tips and warnings.

The Recommendation details your most preferred solution for your case study’s problem.

After identifying and analyzing the problem, your next step is to suggest potential solutions. You did this in the Alternative Courses of Action (ACA) section. Once you’re done writing your ACAs, you need to pick which among these ACAs is the best. The chosen course of action will be the one you’re writing in the recommendation section. 

The Recommendation portion also provides a thorough justification for selecting your most preferred solution. 

Notice how a recommendation in a case study differs from a recommendation in a research paper . In the latter, the recommendation tells your reader some potential studies that can be performed in the future to support your findings or to explore factors that you’re unable to cover. 

Your main goal in writing a case study is not only to understand the case at hand but also to think of a feasible solution. However, there are multiple ways to approach an issue. Since it’s impossible to implement all these solutions at once, you only need to pick the best one. 

The Recommendation portion tells the readers which among the potential solutions is best to implement given the constraints of an organization or business. This section allows you to introduce, defend, and explain this optimal solution. 

How To Write Recommendation in Case Study

You cannot recommend a solution if you are unable to grasp your case study’s issue. Make sure that you’re aware of the problem as well as the viewpoint from which you want to analyze it . 

Once you’ve fully grasped your case study’s problem, it’s time to suggest some feasible solutions to address it. A separate section of your manuscript called the Alternative Courses of Action (ACA) is dedicated to discussing these potential solutions. 

Afterward, you need to evaluate each ACA by identifying its respective advantages and disadvantages. 

After evaluating each proposed ACA, pick the one you’ll recommend to address the problem. All alternatives have their pros and cons so you must use your discretion in picking the best among these ACAs.

To help you decide which ACA to pick, here are some factors to consider:

• Realistic : The organization must have sufficient knowledge, expertise, resources, and manpower to execute the recommended solution. 
• Economical: The recommended solution must be cost-effective.
• Legal: The recommended solution must adhere to applicable laws.
• Ethical: The recommended solution must not have moral repercussions. 
• Timely: The recommended solution can be executed within the expected timeframe. 

You may also use a decision matrix to assist you in picking the best ACA 1 .  This matrix allows you to rank the ACAs based on your criteria. Please refer to our examples in the next section for an example of a Recommendation formed using a decision matrix. 

Provide your justifications for why you recommend your preferred solution. You can also explain why other alternatives are not chosen 2 .  

To help you understand how to make recommendations in a case study, let’s take a look at some examples below.

Case Study Problem : Lemongate Hotel is facing an overwhelming increase in the number of reservations due to a sudden implementation of a Local Government policy that boosts the city’s tourism. Although Lemongate Hotel has a sufficient area to accommodate the influx of tourists, the management is wary of the potential decline in the hotel’s quality of service while striving to meet the sudden increase in reservations. 

Alternative Courses of Action:

• ACA 1: Relax hiring qualifications to employ more hotel employees to ensure that sufficient human resources can provide quality hotel service
• ACA 2: Increase hotel reservation fees and other costs as a response to the influx of tourists demanding hotel accommodation
• ACA 3: Reduce privileges and hotel services enjoyed by each customer so that hotel employees will not be overwhelmed by the increase in accommodations.

Recommendation: 

Upon analysis of the problem, it is recommended to implement ACA 1. Among all suggested ACAs, this option is the easiest to execute with the minimal cost required. It will not also impact potential profits and customers’ satisfaction with hotel service.

Meanwhile, implementing ACA 2 might discourage customers from making reservations due to higher fees and look for other hotels as substitutes. It is also not recommended to do ACA 3 because reducing hotel services and privileges offered to customers might harm the hotel’s public reputation in the long run. 

The first paragraph of our sample recommendation specifies what ACA is best to implement and why.

Meanwhile, the succeeding paragraphs explain that ACA 2 and ACA 3 are not optimal solutions due to some of their limitations and potential negative impacts on the organization. 

Example 2 (with Decision Matrix)

Case Study: Last week, Pristine Footwear released its newest sneakers model for women – “Flightless.” However, the management noticed that “Flightless” had a mediocre sales performance in the previous week. For this reason, “Flightless” might be pulled out in the next few months.  The management must decide on the fate of “Flightless” with Pristine Footwear’s financial performance in mind. 

• ACA 1: Revamp “Flightless” marketing by hiring celebrities/social media influencers to promote the product
• ACA 2: Improve the “Flightless” current model by tweaking some features to fit current style trends
• ACA 3: Sell “Flightless” at a lower price to encourage more customers
• ACA 4: Stop production of “Flightless” after a couple of weeks to cut losses

Decision Matrix

Recommendation

Based on the decision matrix above 3 , the best course of action that Pristine Wear, Inc. must employ is ACA 3 or selling “Flightless” shoes at lower prices to encourage more customers. This solution can be implemented immediately without the need for an excessive amount of financial resources. Since lower prices entice customers to purchase more, “Flightless” sales might perform better given a reduction in its price.

In this example, the recommendation was formed with the help of a decision matrix. Each ACA was given a score of between 1 – 4 for each criterion. Note that the criterion used depends on the priorities of an organization, so there’s no standardized way to make this matrix. 

Meanwhile, the recommendation we’ve made here consists of only one paragraph. Although the matrix already revealed that ACA 3 tops the selection, we still provided a clear explanation of why it is the best. 

• Recommend with persuasion 4 . You may use data and statistics to back up your claim. Another option is to show that your preferred solution fits your theoretical knowledge about the case. For instance, if your recommendation involves reducing prices to entice customers to buy higher quantities of your products, you may invoke the “law of demand” 5 as a theoretical foundation of your recommendation. 
• Be prepared to make an implementation plan. Some case study formats require an implementation plan integrated with your recommendation. Basically, the implementation plan provides a thorough guide on how to execute your chosen solution (e.g., a step-by-step plan with a schedule).
• Manalili, K. (2021 – 2022). Selection of Best Applicant (Unpublished master’s thesis). Bulacan Agricultural State College. Retrieved September 23, 2022, from https://www.studocu.com/ph/document/bulacan-agricultural-state-college/business-administration/case-study-human-rights/19062233.
• How to Analyze a Case Study. (n.d.). Retrieved September 23, 2022, from https://wps.prenhall.com/bp_laudon_essbus_7/48/12303/3149605.cw/content/index.html
• Nguyen, C. (2022, April 13). How to Use a Decision Matrix to Assist Business Decision Making. Retrieved September 23, 2022, from https://venngage.com/blog/decision-matrix/
• Case Study Analysis: Examples + How-to Guide & Writing Tips. (n.d.). Retrieved September 23, 2022, from https://custom-writing.org/blog/great-case-study-analysis
• Hayes, A. (2022, January O8). Law of demand. Retrieved September 23, 2022, from https://www.investopedia.com/terms/l/lawofdemand.asp

Written by Jewel Kyle Fabula

in Career and Education , Juander How

Jewel Kyle Fabula

Jewel Kyle Fabula is a Bachelor of Science in Economics student at the University of the Philippines Diliman. His passion for learning mathematics developed as he competed in some mathematics competitions during his Junior High School years. He loves cats, playing video games, and listening to music.

Browse all articles written by Jewel Kyle Fabula

Copyright Notice

All materials contained on this site are protected by the Republic of the Philippines copyright law and may not be reproduced, distributed, transmitted, displayed, published, or broadcast without the prior written permission of filipiknow.net or in the case of third party materials, the owner of that content. You may not alter or remove any trademark, copyright, or other notice from copies of the content. Be warned that we have already reported and helped terminate several websites and YouTube channels for blatantly stealing our content. If you wish to use filipiknow.net content for commercial purposes, such as for content syndication, etc., please contact us at legal(at)filipiknow(dot)net

Uncovering the Data Science Behind Content Recommendation

Lets deep dive into how Data Science is used for content recommendations, with help of Netflix Case Study.

Introduction

Content recommendation refers to the practice of suggesting content, such as articles, videos, or products, to users based on their interests and behavior. This is an important aspect of many online businesses, as personalized recommendations can increase user engagement and satisfaction, leading to increased revenue and growth.

Data science plays a crucial role in the field of content recommendation. By using techniques such as data mining and machine learning, data scientists can collect and analyze large amounts of data on user behavior, allowing for the creation of personalized recommendations for individual users. In addition, data science is used to improve recommendation algorithms, ensuring that the suggestions made are accurate and relevant.

In this article, we will explore how data science is used in the content recommendation, using the example of Netflix to illustrate the concepts discussed.

How data science is used to understand user behavior

One of the key ways that data science is used in content recommendation is to understand user behavior. By collecting and analyzing data on what users are watching, clicking on, and engaging with, data scientists can gain insights into their interests and preferences.

Data mining techniques are often used to collect this data from various sources, such as website logs, social media, and customer surveys. This data is then cleaned and processed, allowing for the application of machine learning algorithms to uncover patterns and trends.

This information can be used to create personalized recommendations for individual users. For example, if a user has previously watched romantic comedies, a content recommendation system might suggest similar movies or TV shows in the same genre.

In addition, data on user behavior can also be used to improve recommendation algorithms. By analyzing the data, data scientists can evaluate the accuracy and relevance of the recommendations made by the algorithm, and make adjustments as needed to improve its performance.

Few Techniques available for on Recommendation System

There are several techniques that are commonly used in content recommendation systems. These include:

• Collaborative filtering: This approach uses the preferences of similar users to make recommendations. For example, if two users have both watched and enjoyed romantic comedies, the recommendation system might suggest romantic comedies to one of the users based on the other user’s preferences.
• Content-based filtering: This approach uses the characteristics of the content itself to make recommendations. For example, if a user has watched a romantic comedy, the recommendation system might suggest other romantic comedies with similar plot elements or actors.
• Hybrid systems: Some recommendation systems use a combination of collaborative and content-based filtering, leveraging the strengths of both approaches to make more accurate and relevant recommendations.
• Matrix factorization: This is a machine-learning technique that is commonly used in recommendation systems. It involves representing users and items (such as movies or products) as vectors in a high-dimensional space and then using mathematical operations to find relationships between the vectors. This can be used to make predictions about which items a user might be interested in.
• Deep learning: More recently, deep learning techniques have been applied to recommendation systems. These approaches use neural networks to learn complex patterns in the data and make predictions about user preferences.

How data science is used to improve recommendation algorithms

As mentioned earlier, data science is also used to improve recommendation algorithms. This is an important aspect of content recommendation, as accurate and relevant recommendations are crucial for user engagement and satisfaction.

To evaluate and improve recommendation algorithms, data scientists will often use techniques such as A/B testing, where different algorithms are compared to each other to determine which performs best. This may involve randomly dividing users into groups and showing each group different recommendations, then measuring the engagement and satisfaction of each group to determine which algorithm performs best.

In addition, data scientists may use machine learning techniques to fine-tune the parameters of a recommendation algorithm or to develop new algorithms altogether. For example, they might use matrix factorization or deep learning to learn complex patterns in the data and make more accurate predictions about user preferences.

Overall, the goal of using data science to improve recommendation algorithms is to ensure that the suggestions made to users are as accurate and relevant as possible, leading to increased user engagement and satisfaction.

Case study: Netflix

One of the most well-known examples of a content recommendation system is Netflix. The company uses data science extensively to understand user behavior and improve its recommendation algorithms. Netflix collects data on what users are watching and how they are interacting with the platform. This data is then used to create personalized recommendations for individual users. For example, if a user has previously watched romantic comedies, Netflix might suggest similar movies or TV shows in the same genre.

In addition, Netflix also uses data science to improve its recommendation algorithms. The company has developed a number of proprietary algorithms that take into account various factors, such as a user’s viewing history, ratings, and search behavior, to make recommendations. The success of Netflix’s recommendation system has been a key factor in the company’s growth and success. By providing personalized and relevant recommendations, Netflix has been able to increase user engagement and satisfaction, leading to increased revenue and a larger user base.

Overall, the use of data science in content recommendation has played a crucial role in the success of Netflix and other online businesses. By using data mining, machine learning, and other techniques, data scientists can help create personalized and relevant recommendations, leading to increased user engagement and satisfaction.

Few More Examples

There are many companies that use data science for content recommendation. Some examples include:

• Amazon: The online retail giant uses data science to make personalized product recommendations to its users. By analyzing data on user behavior and preferences, Amazon’s recommendation system suggests products that users are likely to be interested in based on their previous purchases and interactions with the platform.
• Spotify: The popular music streaming service uses data science to make personalized music recommendations to its users. By analyzing data on user listening habits, Spotify’s recommendation system suggests songs and playlists that are similar to what the user has previously listened to.
• YouTube: The video-sharing platform uses data science to make personalized video recommendations to its users. By analyzing data on user viewing habits, YouTube’s recommendation system suggests videos that are similar to what the user has previously watched.
• Facebook: The social media giant uses data science to make personalized news feed recommendations to its users. By analyzing data on user interactions and interests, Facebook’s recommendation system suggests posts and articles that are likely to be of interest to the user based on their previous activity on the platform.

Overall, these companies and many others use data science to provide personalized and relevant recommendations to their users, leading to increased engagement and satisfaction.

To summarize, Data science plays a crucial role in the field of content recommendation. By using techniques such as data mining and machine learning, data scientists can collect and analyze large amounts of data on user behavior, allowing for the creation of personalized recommendations for individual users. In addition, data science is used to improve recommendation algorithms, ensuring that the suggestions made are accurate and relevant.

The example of Netflix illustrates the success that can be achieved by using data science in the content recommendation. By providing personalized and relevant recommendations, the company has been able to increase user engagement and satisfaction, leading to increased revenue and growth.

Looking to the future, we can expect to see continued developments in the field of content recommendation and data science. As more data is collected and new techniques are developed, data scientists will continue to play a crucial role in helping online businesses provide personalized and relevant recommendations to their users.

Thanks for reading till the end do share this with your friends. Learn about HOW DATA SCIENCE IS USED IN THE FOOD AND BEVERAGE INDUSTRY

Leave a Comment Cancel Reply

Your email address will not be published. Required fields are marked *

Save my name, email, and website in this browser for the next time I comment.

How to Write a Case Study - All You Wanted to Know

What do you study in your college? If you are a psychology, sociology, or anthropology student, we bet you might be familiar with what a case study is. This research method is used to study a certain person, group, or situation. In this guide from our dissertation writing service , you will learn how to write a case study professionally, from researching to citing sources properly. Also, we will explore different types of case studies and show you examples — so that you won’t have any other questions left.

What Is a Case Study?

A case study is a subcategory of research design which investigates problems and offers solutions. Case studies can range from academic research studies to corporate promotional tools trying to sell an idea—their scope is quite vast.

What Is the Difference Between a Research Paper and a Case Study?

While research papers turn the reader’s attention to a certain problem, case studies go even further. Case study guidelines require students to pay attention to details, examining issues closely and in-depth using different research methods. For example, case studies may be used to examine court cases if you study Law, or a patient's health history if you study Medicine. Case studies are also used in Marketing, which are thorough, empirically supported analysis of a good or service's performance. Well-designed case studies can be valuable for prospective customers as they can identify and solve the potential customers pain point.

Case studies involve a lot of storytelling – they usually examine particular cases for a person or a group of people. This method of research is very helpful, as it is very practical and can give a lot of hands-on information. Most commonly, the length of the case study is about 500-900 words, which is much less than the length of an average research paper.

The structure of a case study is very similar to storytelling. It has a protagonist or main character, which in your case is actually a problem you are trying to solve. You can use the system of 3 Acts to make it a compelling story. It should have an introduction, rising action, a climax where transformation occurs, falling action, and a solution.

Here is a rough formula for you to use in your case study:

Problem (Act I): > Solution (Act II) > Result (Act III) > Conclusion.

Types of Case Studies

The purpose of a case study is to provide detailed reports on an event, an institution, a place, future customers, or pretty much anything. There are a few common types of case study, but the type depends on the topic. The following are the most common domains where case studies are needed:

• Historical case studies are great to learn from. Historical events have a multitude of source info offering different perspectives. There are always modern parallels where these perspectives can be applied, compared, and thoroughly analyzed.
• Problem-oriented case studies are usually used for solving problems. These are often assigned as theoretical situations where you need to immerse yourself in the situation to examine it. Imagine you’re working for a startup and you’ve just noticed a significant flaw in your product’s design. Before taking it to the senior manager, you want to do a comprehensive study on the issue and provide solutions. On a greater scale, problem-oriented case studies are a vital part of relevant socio-economic discussions.
• Cumulative case studies collect information and offer comparisons. In business, case studies are often used to tell people about the value of a product.
• Critical case studies explore the causes and effects of a certain case.
• Illustrative case studies describe certain events, investigating outcomes and lessons learned.

Need a compelling case study? EssayPro has got you covered. Our experts are ready to provide you with detailed, insightful case studies that capture the essence of real-world scenarios. Elevate your academic work with our professional assistance.

Case Study Format

The case study format is typically made up of eight parts:

• Executive Summary. Explain what you will examine in the case study. Write an overview of the field you’re researching. Make a thesis statement and sum up the results of your observation in a maximum of 2 sentences.
• Background. Provide background information and the most relevant facts. Isolate the issues.
• Case Evaluation. Isolate the sections of the study you want to focus on. In it, explain why something is working or is not working.
• Proposed Solutions. Offer realistic ways to solve what isn’t working or how to improve its current condition. Explain why these solutions work by offering testable evidence.
• Conclusion. Summarize the main points from the case evaluations and proposed solutions. 6. Recommendations. Talk about the strategy that you should choose. Explain why this choice is the most appropriate.
• Implementation. Explain how to put the specific strategies into action.
• References. Provide all the citations.

How to Write a Case Study

Let's discover how to write a case study.

Setting Up the Research

When writing a case study, remember that research should always come first. Reading many different sources and analyzing other points of view will help you come up with more creative solutions. You can also conduct an actual interview to thoroughly investigate the customer story that you'll need for your case study. Including all of the necessary research, writing a case study may take some time. The research process involves doing the following:

• Define your objective. Explain the reason why you’re presenting your subject. Figure out where you will feature your case study; whether it is written, on video, shown as an infographic, streamed as a podcast, etc.
• Determine who will be the right candidate for your case study. Get permission, quotes, and other features that will make your case study effective. Get in touch with your candidate to see if they approve of being part of your work. Study that candidate’s situation and note down what caused it.
• Identify which various consequences could result from the situation. Follow these guidelines on how to start a case study: surf the net to find some general information you might find useful.
• Make a list of credible sources and examine them. Seek out important facts and highlight problems. Always write down your ideas and make sure to brainstorm.
• Focus on several key issues – why they exist, and how they impact your research subject. Think of several unique solutions. Draw from class discussions, readings, and personal experience. When writing a case study, focus on the best solution and explore it in depth. After having all your research in place, writing a case study will be easy. You may first want to check the rubric and criteria of your assignment for the correct case study structure.

Read Also: ' WHAT IS A CREDIBLE SOURCES ?'

Although your instructor might be looking at slightly different criteria, every case study rubric essentially has the same standards. Your professor will want you to exhibit 8 different outcomes:

• Correctly identify the concepts, theories, and practices in the discipline.
• Identify the relevant theories and principles associated with the particular study.
• Evaluate legal and ethical principles and apply them to your decision-making.
• Recognize the global importance and contribution of your case.
• Construct a coherent summary and explanation of the study.
• Demonstrate analytical and critical-thinking skills.
• Explain the interrelationships between the environment and nature.
• Integrate theory and practice of the discipline within the analysis.

Need Case Study DONE FAST?

Pick a topic, tell us your requirements and get your paper on time.

Case Study Outline

Let's look at the structure of an outline based on the issue of the alcoholic addiction of 30 people.

Introduction

• Statement of the issue: Alcoholism is a disease rather than a weakness of character.
• Presentation of the problem: Alcoholism is affecting more than 14 million people in the USA, which makes it the third most common mental illness there.
• Explanation of the terms: In the past, alcoholism was commonly referred to as alcohol dependence or alcohol addiction. Alcoholism is now the more severe stage of this addiction in the disorder spectrum.
• Hypotheses: Drinking in excess can lead to the use of other drugs.
• Importance of your story: How the information you present can help people with their addictions.
• Background of the story: Include an explanation of why you chose this topic.
• Presentation of analysis and data: Describe the criteria for choosing 30 candidates, the structure of the interview, and the outcomes.
• Strong argument 1: ex. X% of candidates dealing with anxiety and depression...
• Strong argument 2: ex. X amount of people started drinking by their mid-teens.
• Strong argument 3: ex. X% of respondents’ parents had issues with alcohol.
• Concluding statement: I have researched if alcoholism is a disease and found out that…
• Recommendations: Ways and actions for preventing alcohol use.

Writing a Case Study Draft

After you’ve done your case study research and written the outline, it’s time to focus on the draft. In a draft, you have to develop and write your case study by using: the data which you collected throughout the research, interviews, and the analysis processes that were undertaken. Follow these rules for the draft:

Use Data to Illustrate Key Points in Your Case Study

Even though your case study is a story, it should be based on evidence. Use as much data as possible to illustrate your point. Without the right data, your case study may appear weak and the readers may not be able to relate to your issue as much as they should. Let's see the examples from essay writing service :

‍ With data: Alcoholism is affecting more than 14 million people in the USA, which makes it the third most common mental illness there. Without data: A lot of people suffer from alcoholism in the United States.

Try to include as many credible sources as possible. You may have terms or sources that could be hard for other cultures to understand. If this is the case, you should include them in the appendix or Notes for the Instructor or Professor.

Finalizing the Draft: Checklist

After you finish drafting your case study, polish it up by answering these ‘ask yourself’ questions and think about how to end your case study:

• Check that you follow the correct case study format, also in regards to text formatting.
• Check that your work is consistent with its referencing and citation style.
• Micro-editing — check for grammar and spelling issues.
• Macro-editing — does ‘the big picture’ come across to the reader? Is there enough raw data, such as real-life examples or personal experiences? Have you made your data collection process completely transparent? Does your analysis provide a clear conclusion, allowing for further research and practice?

Problems to avoid:

• Overgeneralization – Do not go into further research that deviates from the main problem.
• Failure to Document Limitations – Just as you have to clearly state the limitations of a general research study, you must describe the specific limitations inherent in the subject of analysis.
• Failure to Extrapolate All Possible Implications – Just as you don't want to over-generalize from your case study findings, you also have to be thorough in the consideration of all possible outcomes or recommendations derived from your findings.

How to Create a Title Page and Cite a Case Study

Let's see how to create an awesome title page.

Your title page depends on the prescribed citation format. The title page should include:

• A title that attracts some attention and describes your study
• The title should have the words “case study” in it
• The title should range between 5-9 words in length
• Your name and contact information
• Your finished paper should be only 500 to 1,500 words in length.With this type of assignment, write effectively and avoid fluff

Here is a template for the APA and MLA format title page:

There are some cases when you need to cite someone else's study in your own one – therefore, you need to master how to cite a case study. A case study is like a research paper when it comes to citations. You can cite it like you cite a book, depending on what style you need.

Citation Example in MLA ‍ Hill, Linda, Tarun Khanna, and Emily A. Stecker. HCL Technologies. Boston: Harvard Business Publishing, 2008. Print.

Citation Example in APA ‍ Hill, L., Khanna, T., & Stecker, E. A. (2008). HCL Technologies. Boston: Harvard Business Publishing.

Citation Example in Chicago Hill, Linda, Tarun Khanna, and Emily A. Stecker. HCL Technologies.

Case Study Examples

To give you an idea of a professional case study example, we gathered and linked some below.

Eastman Kodak Case Study

Case Study Example: Audi Trains Mexican Autoworkers in Germany

To conclude, a case study is one of the best methods of getting an overview of what happened to a person, a group, or a situation in practice. It allows you to have an in-depth glance at the real-life problems that businesses, healthcare industry, criminal justice, etc. may face. This insight helps us look at such situations in a different light. This is because we see scenarios that we otherwise would not, without necessarily being there. If you need custom essays , try our research paper writing services .

Get Help Form Qualified Writers

Crafting a case study is not easy. You might want to write one of high quality, but you don’t have the time or expertise. If you’re having trouble with your case study, help with essay request - we'll help. EssayPro writers have read and written countless case studies and are experts in endless disciplines. Request essay writing, editing, or proofreading assistance from our custom case study writing service , and all of your worries will be gone.

Don't Know Where to Start?

Crafting a case study is not easy. You might want to write one of high quality, but you don’t have the time or expertise. Request ' write my case study ' assistance from our service.

What Is A Case Study?

How to cite a case study in apa, how to write a case study.

Daniel Parker

is a seasoned educational writer focusing on scholarship guidance, research papers, and various forms of academic essays including reflective and narrative essays. His expertise also extends to detailed case studies. A scholar with a background in English Literature and Education, Daniel’s work on EssayPro blog aims to support students in achieving academic excellence and securing scholarships. His hobbies include reading classic literature and participating in academic forums.

is an expert in nursing and healthcare, with a strong background in history, law, and literature. Holding advanced degrees in nursing and public health, his analytical approach and comprehensive knowledge help students navigate complex topics. On EssayPro blog, Adam provides insightful articles on everything from historical analysis to the intricacies of healthcare policies. In his downtime, he enjoys historical documentaries and volunteering at local clinics.

An explainable content-based approach for recommender systems: a case study in journal recommendation for paper submission

• Open access
• Published: 06 June 2024
• Volume 34 , pages 1431–1465, ( 2024 )

Cite this article

You have full access to this open access article

• Luis M. de Campos 1   na1 ,
• Juan M. Fernández-Luna 1   na1 &
• Juan F. Huete 1   na1  

646 Accesses

Explore all metrics

Explainable artificial intelligence is becoming increasingly important in new artificial intelligence developments since it enables users to understand and consequently trust system output. In the field of recommender systems, explanation is necessary not only for such understanding and trust but also because if users understand why the system is making certain suggestions, they are more likely to consume the recommended product. This paper proposes a novel approach for explaining content-based recommender systems by specifically focusing on publication venue recommendation. In this problem, the authors of a new research paper receive recommendations about possible journals (or other publication venues) to which they could submit their article based on content similarity, while the recommender system simultaneously explains its decisions. The proposed explanation ecosystem is based on various elements that support the explanation (topics, related articles, relevant terms, etc.) and is fully integrated with the underlying recommendation model. The proposed method is evaluated through a user study in the biomedical field, where transparency, satisfaction, trust, and scrutability are assessed. The obtained results suggest that the proposed approach is effective and useful for explaining the output of the recommender system to users.

Similar content being viewed by others

Explainable Recommender Systems: From Theory to Practice

Metrics for Evaluating Explainable Recommender Systems

Personalized Review-Oriented Explanations for Recommender Systems

Explore related subjects.

• Artificial Intelligence

Avoid common mistakes on your manuscript.

1 Introduction

Recommender systems (RS), a form of artificial intelligence (AI), offer personalized suggestions based on users’ preferences and behaviors, widely utilized in online retail, streaming services, and social media platforms. Recognizing the significance of explainable AI both in a general context (Minh et al. 2022 ; Li et al. 2023 ; Barredo-Arrieta et al. 2020 and specifically within RS (Pavitha et al. 2022 ; Zhang and Chen 2020 ), the provision of explanations for recommended content serves multiple purposes. It not only enhances user trust but also improves transparency, allowing users to comprehend the reasoning behind recommendations, mitigates biases arising from data or algorithms, reinforces users’ reliance on the system’s judgments, and fosters increased engagement (Tintarev and Masthoff 2007 ; Zhang and Chen 2020 ). In this paper, we present an explanation approach tailored to enhance user understanding of RS recommendations in a venue recommendation problem, whereby a researcher with a recently written paper needs to decide to which journal the paper should be submitted.

When the RS is provided with the title and abstract of a new article, it suggests a group of appropriate venues for publication. The system uses a content-based RS (CBRS) where each journal included in the RS is represented by various textual subprofiles which group together every article published in the journal on the same topic. The input text is, therefore, matched against the journal subprofiles, and the user is presented with a sorted list of recommended journals in decreasing order according to their associated scores, which represent a type of aggregation of the subprofile scores.

Nevertheless, once the RS has recommended the most suitable journals according to the submitted title and abstract for the user to consider, users might not understand why certain suggestions were made and this could well result in loss of confidence in the RS. The proposed explanation approach is based on different explanation elements that generate information from different components from the CBRS in order to justify the recommendations (confidence in the recommendation, global topic word clouds, similar articles to the target one published in the same journal, specific topic word clouds covered by the journal and certain highlighted words from the target title and abstract). The proposed method is based only on content, i.e., the text of the articles published in their corresponding journals. Although other bibliometric information (such as impact factor, quartiles, co-authorship, etc.) might be used, the scope of this paper only focuses on text so as to measure the feasibility of this approach without external information.

This paper not only presents the design of this integral explanation scheme but also the results of a user study in order to determine how selected biomedical experts (the context of this evaluation) viewed the explanation tools presented and how useful they considered them to be.

The paper makes the following contributions:

A novel approach for explaining CBRS which is totally integrated in the recommendation model particularized in the journal recommendation problem

Verification of the transparency, satisfaction, trust, and scrutability of the proposal by means of a user study

A comparison of the proposed method with those found in the state of the art

The remainder of the paper is organized as follows: Sect.  2 examines related work on explanations in recommender systems; Sect.  3 provides a brief overview of the specific system for recommending scientific journals for which we have developed explanation facilities; in Sect.  4 , we describe different explanation elements developed; then, Sect.  5 presents the user study designed to evaluate the performance of the explainable recommender system from different perspectives; and finally, Sect.  6 outlines our conclusions and future lines of research.

2 State of the art

One of the major contributions dealing with explanation in RSs is the work by Tintarev and Masthoff ( 2007 ). They justify the implementation of good explanations since they can “help inspire user trust and loyalty, increase satisfaction, make it quicker and easier for users to find what they want, and persuade them to try or purchase a recommended item.” Explainable RS (XRS) address the problem of why such items are recommended (Zhang and Chen 2020 ). Tintarev and Masthoff established seven benefits of the explanation: transparency (explain how the RS works), scrutability (users are able to express whether the RS is correct or not), trust (increase user confidence in the system), effectiveness (help the user make good choices), persuasiveness (make the user perform an action), efficiency (make decisions faster), and satisfaction (fulfillment of the user’s needs). Two additional purposes might be education (learn something from the system) and debugging (identify problems in the RS) (Jannach et al. 2019 ). These elements are valid when designing RS explanation features and reflect the dimensions on which these should be based.

2.1 Classification of explainable recommender systems

The explanations provided by an RS could be classified into two models (Zhang and Chen 2020 ) depending on the interpretability of the explanations: while the model-intrinsic approach offers direct transparency for the RS decision, the model-agnostic one needs to create explanations after the decisions are taken. In this last case, the RS is a black box and explanation must be built on the top of it with all available resources. An example of this type is what Shmaryahu et al. ( 2020 ) call post hoc explanation. This is applied when the recommendation engine is based on a complex model with a low explainability level. A transparent model with a high explainability degree is then used to explain the recommendation. A second example can be found in another article (Iferroudjene et al. 2023 ), whereby the authors create an explanation context based on subgroup discovery on top of a top-n RS to identify active data for the recommendation. Moreover, Papadimitriou et al. ( 2012 ) established an alternative categorization of the XRSs according to the resources used for explanation: human style, based on recommendations of similar users; item style, based on suggestions made to the user on similar items; and feature style, supported by features of an item which were previously considered interesting by a user. In addition, another classification of the explanation is presented in Radensky et al. ( 2022 ), where XRSs might explain a specific item recommendation (local), or in a model-based approach, global, which explains how recommendations are generally generated. Tintarev and Masthoff ( 2012 ) categorized the explanations as personalized or non-personalized according to whether the explanations are specific for each user or suit everyone. Finally, a recent categorization groups the explanation-based models into those where the recommendation model and the explanation function are separated, and the so-called recommendation-by-explanation models, where both actions are integrated (Rana et al. 2022 ). In these latter models, the user does not receive decreasingly sorted recommendations in terms of the confidence of the recommendation model, as it normally does, but rather explanations which are decreasingly sorted according to their quality.

2.2 Explanation styles

Regarding the explanation styles, Tintarev and Masthoff in Tintarev and Masthoff ( 2015 ) determine that there are different explanation styles according to the recommending model used to generate the recommendations: case, content, collaborative, demographic, knowledge, and utility based. Correspondingly, Nunes and Jannach ( 2017 ) identifies four types of explanation content: user preferences or user-provided or user-gathered input; inference process, extracted from the recommendation algorithm itself; background and complementary information such as additional information outside the user’s context and based on their features (list of features and their advantages and disadvantages, and the most relevant explanation features).

Focusing on previously published explanation types, according to the taxonomy of explanations proposed by Nunes and Jannach ( 2017 ), explanations could be classified according to their presentation format into natural language-based explanations (e.g., predefined templates that are instantiated before explanation or structured language) and multimedia-based explanations (visualization in the form of graphs or other media formats).

2.3 Content-based explanations

Since the explanation approach presented in this paper is based on a content-based RS (CBRS), in this section we proceed to present a number of examples. In Verbert et al. ( 2013 ), an explanation approach is presented based on content and tag recommendation in the context of suggesting papers and talks from conferences. The first is supported by a user’s profile from the text from the papers that the user has read, and the recommendation is performed by similarity computation. The second, however, uses tags that are assigned to conference talks. Once again, a matching is carried out between the tags of interest to the user and those from the talks. Similar users are also recommended according to profile closeness, and all the recommended information is shown as a clustermap, combining recommended users, tags, talks, and papers. Cardoso et al. ( 2019 ) present IntersectionExplorer, a recommender system with explanation features in the context of conference paper recommendation. In this, they combine personal, social, and content relevance to perform the recommendation and enable multiple item sets from these relevance perspectives to be explored by means of a user interface. The recommender models applied are tag, bookmark, and bibliography based, with all of them using textual or tag content. Millecamp et al. ( 2019 ) designed an explanation for a CBRS, where profiles and items were represented by features. After computing the corresponding similarity, the explanation was given by a visual element that summarized the values of the user’s features in terms of an interval, the exact values of the features for the recommended items, and a 5-point Likert scale to show the similarity of the recommended item with the user’s profile. In Hernandez-Bocanegra and Ziegler ( 2020 ), an explanation is built on a model based on the user’s profile and its latent features are matched against the latent features of the items and complemented with features representing sentiment information to be extracted from item reviews. The explanation elements were bar charts and tables to explain in terms of the features why the recommended items were selected.

In Sullivan et al. ( 2019 ), within the context of online news recommendation based on the user profiles storing topics, entities, and tags extracted from the read news articles, Sullivan et al. show new topics which might be of interest to the users and associated news. They use three explainability levels: visualization of the distribution of monthly read topics (user’s reading behavior); visualization of that same distribution by the average user (contextualization with the community’s behavior); and new recommended topics explained with the degree of relationships between the user’s topics and new ones (exploitation of the user’s profile). In Polleti and Cozman ( 2019 ), Polleti and Cozman proposed an explanation approach based on topic modeling to explain CBRS (this is, in fact, a model-agnostic method that could be used with any recommendation model): latent topics are extracted from the textual representation of the items and users’ profiles. After obtaining the list of nearest items for a given user’s profile, the common topics between a recommended item and the corresponding profile are then shown in order to justify the recommendation. In Pérez-Núñez et al. ( 2022 ), Pérez-Núñez et al. build a model to recommend and explain textual restaurant reviews written by other TripAdvisor users. Beginning with keywords from the textual reviews which are encoded using BoW, and given a user’s information need, they use a classifier (multinomial logistic regression) to obtain recommendations. The explanation is a tag cloud generated by means of the common terms found in the user’s query and the recommended items.

The paper Tsai and Brusilovsky ( 2019 ), in which the authors present an explainable hybrid recommended system in the context of conference recommendation, integrates several recommendation models based on content (keywords and topic similarities, respectively), social (co-authorship), and demographic information. They created five explanation elements, each generated from a given recommendation model (Venn diagrams containing tag clouds, topic similarities, co-authorship graphs, interest similarity, and geographic distance). Along the same lines, Louki et al. ( 2020 ) shows various explanation mechanisms for a hybrid RS. In terms of the content-based component, the user’s profiles store tags and keywords, and tag and content-based similarities between the profiles and items are incorporated into the hybrid recommendation. The general explanation comprises single explanations, each obtained from every model involved in the recommendation. For CBRS, therefore, they explain the recommendation of an item since it contains similar tags to those found in the profile and there are common tags between them.

In conclusion, in CBRS-based explanation the available information usually used to explain the recommendation is the user’s profile and comprises keywords, tags, or features. Additionally, latent topics extracted from the profiles and item texts are also considered. In terms of the explanation type, most approaches show in one way or another how the item covers the users’ profiles.

2.4 Explanation elements

Going one step further, and independently of the previous explanation styles, a number of already published common types of explanation elements could be used to explain the recommendation: histograms, tables, or pie chart of ratings which show the rating distribution of similar users or ratings of similar products (Daher et al. 2017 ; Zhang and Chen 2020 ; Felfernig et al. 2021 ; Bilgic and Mooney 2015 ; Jannach et al. 2019 ; Pérez-Núñez et al. 2022 ); personalized or non-personalized tag clouds of the keywords that describe the recommended items (Gedikli et al. 2011 ; Chen 2013 ; Daher et al. 2017 ; Felfernig et al. 2021 ; Pérez-Núñez et al. 2022 ); common features or aspects between users and recommended items (Zhang and Chen 2020 ; Zhang et al. 2022 ; Vig et al. 2009 ; Millecamp et al. 2019 ; Pérez-Núñez et al. 2022 ) in the form of Venn diagrams (Tsai and Brusilovsky 2019 ); confidence percentage of the RS prediction (Daher et al. 2017 ); textual descriptions to show the reason for the recommendations (Daher et al. 2017 ); features detected in item descriptions and highlighted (Li et al. 2021 ), radar charts, which include the degree of various features for the recommended items (Daher et al. 2017 ; Zhang and Chen 2020 ; Felfernig et al. 2021 ; Tsai and Brusilovsky 2019 ; elaborated) users’ opinions on the recommended items (Zhang and Chen 2020 ; Hernandez-Bocanegra et al. 2020 ) and their graphical representation (Hernandez-Bocanegra and Ziegler 2020 ); generated natural language (Lully et al. 2018 ); keyword explanation, identifying the terms common to the recommended item and the user’s profile (Bilgic and Mooney 2015 ); list of similar items with the corresponding user’s rating and their impact on the recommendation (Bilgic and Mooney 2015 ) or similar items according to the user’s preferences and their neighbors’ (Shmaryahu et al. 2020 ); graphs, in the form of co-authorship (Tsai and Brusilovsky 2019 ), or bipartite ones containing users and items (Afchar et al. 2022 ) or items connected to their underlying topics (Polleti and Cozman 2019 ); or simply the formula to compute the relevance of a recommended item (Tsai and Brusilovsky 2019 ). As shown, there is a wide variety of explanation elements which depend on the available information and the recommending model.

2.5 Evaluation

The impact of the explanation, meanwhile, can be evaluated directly by means of a user study based on recruited users and a given task (Zhang and Chen 2020 ). The results of such a study are obtained by analyzing questionnaires completed by the users once they have finished the study. Two further methods have also been published (Vultureanu-Albisi and Badica 2022 ), one of which is online (real interaction with an RS) and the other is offline (where no users are considered in the evaluation). Another alternative is to directly measure the impact in the users’ performance in a real environment, i.e., how accurate the decisions are, or how fast they are made (Jesus et al. 2021 ).

When designing an evaluation, since the designer must choose between some of the aims discussed at the beginning of this section as some might be mutually incompatible, any evaluation needs not only to identify the aim being investigated but also to employ suitable metrics (Tintarev and Masthoff 2012 ), mostly a questionnaire with the appropriate questions: for example, a user study is employed to evaluate persuasiveness and usefulness in Sato et al. ( 2019 ) where two specific questions were asked the users in a questionnaire (7-points Likert scale): For persuasiveness, The explanation is convincing and The explanation triggers ; for usefulness, The explanation is useful for choice and The explanation is easy to understand . Persuasiveness is also evaluated in a user study in Louki et al. ( 2020 ), using the question This explanation for the recommendation is convincing . Effectiveness is mainly evaluated in Yao et al. ( 2022 ) with the question Does the explanation help you decide whether you want to watch this recommended movie? . Satisfaction is considered in Ferwerda et al. ( 2018 ) with the question I found the programs that I chose to watch good . In the study by Shulner-Tal et al. ( 2022 ), the System Causability Scale (based on system usability criteria) was utilized. Participants were instructed to assign scores to ten statements, and these scores were subsequently averaged for the comparison of different explanation styles.

3 Overview of the journal recommender system

In this section, we briefly describe the recommender system for which we have designed an explanation module (further details can be found in de Campos et al. 2022 ). This is a content-based system which is specifically devoted to recommending publication venues, and more specifically scientific journals, which might be suitable when attempting to publish a given article based on its content. Although the model may be used in any knowledge domain, the current version has been trained with a collection of articles in the biomedical domain extracted from PubMed/Scopus (Albusac et al. 2018 ).

Our system is based on a representation of each journal through a set of homogeneous thematic subprofiles, covering different topics considered within the journal. To achieve this, we begin with a document collection containing articles published in all the journals under consideration. Subsequently, we employ a text clustering algorithm, specifically K-means, Footnote 1 to group these articles into K clusters Footnote 2 of thematically similar content. The clustering algorithm takes into account only the title, abstract, and keywords of each article. Each cluster, or more precisely, the terms within its constituent articles, represents a global topic within this collection.

We then analyze how each journal addresses each topic. For a given journal, we group together all its articles which belong to the same cluster, creating the corresponding journal subprofile. If no articles from a particular journal belong to a given cluster, that journal’s subprofile for that cluster remains empty. The text of all these topically similar articles is then concatenated into a single macro-document, which represents the subprofile of the journal associated with the corresponding cluster/topic.

We, therefore, have a set of at most K documents associated with each journal. All of these documents from all the journals form a collection of subprofiles, which is subsequently indexed (we use the Lucene library Footnote 3 ) to be used by an information retrieval system (IRS). We employ a language model with Jelinek–Mercer smoothing for this purpose.

Given a target article for which we want a journal recommendation for publication, we use its text as the query for the IRS. A list of the top-h subprofiles Footnote 4 is then retrieved, together with the corresponding scoring values scr ( j ,  t ) indicating the relevance for the candidate article of those papers published in the journal j under the topic t . In other words, we compute a similarity degree between the candidate article and the way each journal j approaches each topic t . It is worth noting that this ranking shows a many-to-many relationship between topics and journals. While the same topic t might appear in various relevant journals, a given journal j may also encompass different topics relevant to the query.

To generate a list of recommended journals, we transform this subprofile ranking into a journal ranking using a fusion algorithm (de Campos et al. 2017 ). This algorithm aggregates the scores scr ( j ,  t ) from all subprofiles retrieved from the same journal j , applying logarithmic penalization to account for their ranking position. This ensures that lower ranked subprofiles have a less influence on the final recommendation. The rationale is that if, for example, the candidate article primarily pertains to a topic \(t_1\) but also touches on a topic \(t_2\) to a lesser extent, a journal that covers both of these topics appears more relevant than a journal covering only one of them. The top n journals Footnote 5 from this newly generated ranking are then recommended to the user, as depicted in Fig.  1 .

Text of the target article and list of recommended journals

As we have seen, the recommendation process is based on information about topics (high level) as well as terms (low level) and their similarity to the query. However, the current system does not utilize any of the internal information it manages (such as what topics are discovered by the clustering algorithm, what articles and terms form the subprofiles associated with each journal, and what terms from these subprofiles match with those of the target article) to explain its recommendations. Instead, it simply provides an ordered list of recommended journals. Therefore, we attempt to examine and exploit these internal processes that the system carries out to generate meaningful explanations for its recommendations.

4 Description of the explanation elements

As stated in the previous section, once the user has entered the information about a target article (title, abstract and keywords), the system retrieves a ranked list of journals and there is the possibility of accessing different explanation elements. We have proposed various types of explanation, not only to test which of these are preferred by the users but also because, according to a number of previous studies (Louki et al. 2020 ; Papadimitriou et al. 2012 ; Tsai and Brusilovsky 2019 ), the combination of different types of explanations can be positive.

It should be noted that although the explanation elements considered are adapted to the specific type of application considered, namely recommending scientific journals, as long as the base recommender system can be applied to other domains (such as, for example, expert finding), these elements can also be easily adapted.

Also, it is interesting to highlight that provided that the recommendation system is capable of offering a journal recommendation, an explanation for each one of them can also be generated as these explanations come from information that the RS uses to generate the recommendation. If the recommendation is bad, the explanation could be useful for the user to detect the inability of the RS for creating a quality recommendation.

In our case, we consider two levels of explanation: a global level that attempts to provide a general idea of why the entire set of journals was recommended, and a local level, where the explanations focus on each specific recommended journal.

4.1 Global explanation: why is this ranking obtained?

The two explanation elements (EE) considered in this level can be articulated in the following way.

4.1.1 EE1: ranking confidence

The original output of our recommender system was a list of journals, as depicted in Fig.  1 . However, from the end user’s perspective, it can be challenging to discern why a particular journal is ranked higher or lower than others and to what extent. Knowing such information can help users to make informed decisions regarding the most suitable journal for submitting the paper, ultimately enhancing the recommending experience.

Therefore, the first explanation element consists of simply showing a numerical score associated with every recommended journal, representing the matching degree between the target article and each journal, and reflecting system confidence in its recommendations (see Fig.  2 , where in the caption of this figure we include the explanatory text given to the users). Opting for normalized scores (expressed as percentages) rather than raw scores enables consistent comparisons across various recommendations, irrespective of the retrieval algorithm or the specific target article. This choice enhances comparability and interpretation.

In particular, the used scores range from 0 to 100%, with higher values signifying the journal’s greater suitability for publishing our article. They are obtained by dividing the raw score, a topic-based similarity degree, by the maximum score within the set of retrieved journals. Thus, from Fig.  2 , it becomes evident that the topics covered in our article exhibit (approximately) twice the degree of similarity to the topics in the journal Pediatrics in comparison with the topics discussed in the journal Developmental Medicine and Child Neurology .

EE1: matching degrees between the recommended journals and the target article in Fig.  1 . The given explanation text is: “Top 10 recommended journals and the relative confidence of the system in each recommendation (the top journal always receives 100% confidence)”

In previously published related work (Bilgic and Mooney 2015 ), the authors compute an influence degree of each item on the recommendation. This value is computed by removing the item from the training set and then recomputing the recommendation value for all the test items and finally considering the difference in scores with and without the item. Their experimental results show that the explanation based on this influence degree was very effective in comparison with the other two tested explanations. In Daher et al. ( 2017 ), they assign to each item the percentage of correct predictions in a system where the users can rate the prediction and feed the RS with that information. Although proved as useful, the problem of needing a sufficient number of predictions makes it not very operative. Our approach directly uses the (normalized) scores obtained from the CBRS to build a confidence percentage.

4.1.2 EE2: related topics

Topics play a crucial role in our system, as they enable more effective recommendations (de Campos et al. 2022 ). However, these topics were automatically derived from the entire corpus of published papers. As a consequence, there might be a gap between researchers’ intuitions and the learned topics. In order to bridge this gap, another aspect of the explanation entails identifying the core themes discussed in the target article and scrutinizing how they are treated in the recommended journals. In order to rapidly and intuitively outline the essence of a topic, we selected the most significant terms (in a word cloud and/or a sorted list of terms), a widely employed technique in topic modeling (Aletras et al. 2017 ; Chi et al. 2019 ).

Continuing with our target article example (Fig.  1 ), the two word clouds in Fig.  3 illustrate how this paper can be approached from two distinct angles: one related to “ children health ” (Topic 56) and the other linked to “ controlled interventions studies ” (Topic 60). We also indicate in the caption the explanatory text presented to the user. Through the examination of these high-level topics, users can gain a more comprehensive understanding of the suitability of the recommended journals. Thus, in case these topics do not align with our expectations, we may have reservations about trusting the provided recommendations.

For the purpose of showing the related topics, two independent tasks must be undertaken. The first one is related to the approach used to determine the topic contents, and the second one involves establishing the number of relevant topics to display, ensuring it does not overwhelm the users.

EE2: word clouds of related topics. The given explanation text is: “ Within the recommended journals, these two topics are the more relevant to your research. The word clouds showcase the most commonly used words under each respective topic. ” From these word clouds, researchers can see that papers focused on studies of children health as well as controlled interventions which have been published in the recommended journals are the most relevant to their research

4.1.3 Showing topics descriptions

In our case, topics has been learned from the entire collection of papers, grouping individual papers with a shared subject into the same cluster. In a broader sense, each topic is formed by the amalgamation of all the papers within the same cluster (i.e., all the journal subprofiles associated with this topic). For instance, Topic 56 and Topic 60 (in Fig.  3 ) are addressed in a total of 410 and 257 journals, respectively, with quite different scopes.

Thus, selecting the most significant terms from this amalgamation to describe a topic, i.e., utilizing the entire collection, might not contribute effectively to the understanding of our recommendations. This is because the resulting description can lead to a significant deviation from the context of our target research paper, making it challenging for the user to discern its relevance. In our running example, Topic 56 is predominantly characterized by terms related to children asthma treatment , as it is the prevalent health concern in a substantial number of papers published under this topic. Therefore, presenting such information would lead to confusion in understanding our explanations, as our paper is not related to asthma.

In order to overcome this problem, we specifically restrict the topic description to those journal subprofiles that are highly relevant to our paper. These subprofiles are those occupying the top positions in the ranking. This deliberate selection ensures that the obtained description is much more focused on the specificity of the target article, thereby enhancing its comprehensibility for the user.

We can say that the generated descriptions are personalized since they vary with the target research paper. Also, from a computational point of view, it is important to highlight the fact that we are referring to a relatively small number of journals discussing the topic, typically in the range of tens.

After choosing the journal subprofiles employed as the source text to build the word clouds, we provide a brief overview of the process used to select the most significant terms. Firstly, we utilize the spaCy library for entity recognition, identifying word groupings that exhibit strong relationships (in a general context, “White House” would be considered an entity) and treating them as a single unit. Any words not recognized as entities are considered separately. The resulting word stream is lemmatized, excluding the stop words, which are removed. Lemmatization is performed using the WordNetLemmatizer library in Python, taking into account their respective parts of speech in the sentence, determined using the NLTK POS tagger, also in Python. The resulting lemmas are then ready for counting their occurrences in the text stream to create the word clouds (with the top 50 lemmas being selected) or the list of terms (we show the top 15 Footnote 6 ).

4.1.4 Selecting the number of topics

Since it is rare for an article to be related to many high-level topics, only the most appropriate ones should be selected to explain the recommendation: sometimes, a single topic can fully explain the article, whereas other times, the inclusion of multiple topics can provide helpful explanations. However, for a given target article the topic ranking typically includes dozens of topics, being many of then marginally relevant. Footnote 7 As a consequence, an automated method for identifying the optimal number of topics is necessary.

For this purpose, we use a method which involves selecting the appropriate number of items from a set of available items. This methodology was proposed in (de Campos et al. 2021 ) Footnote 8 where the input is a set of m items of any type and the importance of each is represented by a weight \([w_1, \ldots , w_m]\) ; the output is the most important items.

In our case, the topics are the items and the weights are obtained by aggregating, for each topic t , the scores across all relevant journals (subprofiles) related to it, expressed as \(w_t = \sum _j scr(j,t)\) .

The methodology is based on ranking the topics in decreasing order of their weights and comparing, through a similarity or a distance measure, the probability distribution obtained by normalizing the topic weights with the vectors \((1,0,0,\ldots ,0)\) , \((1,1,0,\ldots ,0)\) , \((1,1,1,\ldots ,0), \ldots , (1,1,1,\ldots ,1)\) . Each vector represents the decision of selecting the first \(k=1,2,\ldots \) topics with the greatest weight. The optimal decision is the one that optimizes the distance or similarity measure being used. We denote this function as TopicSelection \(([w_{t_1}, \ldots , w_{t_m}], measure)\) . It should be noted that the number of topics selected is not constant and relies on their individual weights. Many distance or similarity measures were considered in (de Campos et al. 2021 ), although all of these converged into only five selection strategies, ranging from the most restrictive Euclidean measure which always chooses the top ranked topic to the Overlap measure that selects all the topics with nonzero weights. Intermediate strategies employ the Dice, Sorensen, or Cosine measure.

We conducted an offline experiment to determine the appropriate metric for this selection problem using a set of 32,864 test queries from our dataset (see Sect.  5.2 ). We counted the number of topics selected for each case and found that on average, the Overlap measure shows 11.96 topics, which is obviously excessive. Using the proposed methodology, we were able to select an average of 1.14, 1.69, and 2.85 topics using the Dice, Sorensen, and Cosine measures, respectively. The average number of topics selected using Dice was too restrictive and almost always (87% of the times), only one topic was displayed. The Cosine measure, on the other hand, often selected a high number of topics (24.4% of the times select four or more topics, with a maximum of 35 topics), which could be excessive. Sorensen measure represents a compromise (96.5% of the times shows one to three topics, with a maximum of eight different topics). Thus, given that papers can relate to multiple topics, but not an excessive number, we decided to use the Sorensen measure for our explanations, i.e., TopicSelection ( \([w_{t_1}, \ldots , w_{t_m}]\) , Sorensen) is applied.

To conclude this section, we relate our approach to those in the literature. As previously mentioned, a number of different tag cloud approaches have previously been described and these include (Gedikli et al. 2011 ), where the authors present a basic approach where a tag cloud is created directly from the words in the item description and their number of occurrences. In addition, they show a second alternative, called the personalized tag cloud, where the users express their interests in the tags and the tag-based diagram shows how the items cover such interests. The authors of this paper founded the use of tag cloud in the fact that “explanations based on tag clouds are not only well accepted by the users but can also help to improve the efficiency and effectiveness of the explanation process.” An aggregated view of several users in the context of group recommendation in tag cloud form is presented in Felfernig et al. ( 2021 ), where the graphic contained not only the tags but also the indication of which users preferred each tag, resulting in a very interesting way of explaining. Pérez-Núñez et al. ( 2022 ) propose a tag cloud where the importance of the tags that characterize a product is not the typical term frequency but is learned by means of a deep learning process. In Chen ( 2013 ), based on a collaborative filtering RS where the items contain a textual representation, the tag clouds are colored according to terms found in the user’s positive, negative, or neutral reviews. The suggested method can enhance the acceptance rate of recommendations and enhance user satisfaction.

The word clouds for our second explanation element are generated in a totally different way to these approaches. Firstly, we find ourselves faced with a CBRS context; secondly, we could say that the recommendations are personalized, but not in the way proposed in Gedikli et al. ( 2011 ) but in terms of the user’s input (the article submitted to the RS) as the recommendations are first generated from the most relevant subprofiles in the ranking by means of a process of entity detection and then from a word frequency count to determine word size.

4.2 Local explanations: why is a particular journal recommended?

The system also offers more specific explanations about each of the recommended journals. By clicking on the magnifying glass to the right of the journal title (see Fig.  2 ), a detailed explanation for each journal can be found.

4.2.1 EE3: journal-related articles

The first specific explanation (Element 3) for a journal is a list of up to three articles published in the journal which are most similar to the target article (Fig.  4 shows the list associated with the first and third recommended journals for the target article used in Fig.  1 ).

Each article has associated a traffic light, where the colors means the degree of similarity with respect to the target article: green, very similar; orange, medium similarity; and red, low similarity.

In this way, users can observe how their target articles are similar in content to other articles previously published in the journal, and this provides a different, more specific reason for the journal recommendation.

EE3 Explanation text: “ List of articles published for the recommended journal, which are similar to the target article. The colors of the traffic lights represent the degree of similarity with respect to this target article (green very similar; orange, medium similarity; red, low similarity) . This figure shows EE3 explanations for the first (top) and third (bottom) recommended journals, Pediatrics and Early Human Development , respectively, which are similar to the target article in Fig.  1

In order to find the set of related articles, we used an auxiliary index to index each paper separately rather than topic-based subprofiles. We submitted the same query (title, abstract and keywords) and obtained a ranked list of related articles using the same similarity measure, i.e., the Jelineck–Mercer language model.

For each recommended journal, we decided to show up to three articles (the most relevant ones published in the journal), but only if they belong to the 50 most similar ones. This is because we believe that articles beyond this threshold may not be sufficiently similar to the target paper. It is worth noting that as a result of this restriction, it is possible that zero, one, or two articles may be displayed for a given journal. This may be an indication of a possibly inappropriate journal recommendation.

In order to help users quickly identify the relevance of the displayed articles, we have color-coded them: articles with a high relevance (ranked in the top 10) are displayed in green, those with a medium relevance (ranked between 11 and 25) are displayed in orange, and those with a low relevance (ranked below 25) are displayed in red. The position of each article in the ranking is also shown.

This explanation element is related to what Bilgic and Mooney ( 2015 ) call the Influence Style Explanation: in the context of the explanation, they show the training items with the greatest impact on the recommendation and their corresponding user’s ratings, as well as a score representing their influence on the recommendation. It is a class of similar items to the one recommended but based on users’ previous experience. Shmaryahu et al. ( 2020 ) also justify the recommendation showing some similar and previously rated items to the recommended one, although this technique is not the best one for the users in their study. In our case, this explanation element does not offer previously rated items but similar articles to the target one published in the recommended journal, based exclusively on content similarity. Moreover, for this explanation element 3, the list of similar items does not have any direct impact on the recommendation and is only a mere explanation in contrast with how the similar items are used in Bilgic and Mooney ( 2015 ) and Shmaryahu et al. ( 2020 ).

4.2.2 EE4: journal-related topics

Another explanation element, specific for each recommended journal, consists of a set of word clouds that depict the primary topics and their coverage within the journal. While we employ a similar representation technique as in EE2, utilizing word clouds in an abstract manner, there are several distinctions. Firstly, we exclusively consider relevant subprofiles from the specific journal as input so that the topics selected (the most representatives for this journal) can be different from those in the global explanation EE2. Moreover, in cases where the same topic is selected, variations may emerge in how that topic is specifically addressed within the journal. This distinction is exemplified in Fig.  5 , displaying the word cloud representation of Topic 56 in Developmental Medicine and Child Neurology , the sixth recommended journal. Globally, Topic 56 was associated with child health research in EE2 (see Fig.  3 ). However, within the recommended journal, its focus is narrowed to issues related to motor function and cerebral palsy in children . Such insights can be valuable for users in determining the potential relevance of this journal for publishing their research.

Another difference lies in the methodology for determining the number of topics presented to the user. To address this, we employ the same topic selection algorithm as previously discussed. However, in this scenario, the weights are determined by the raw scores \(w_{tj} = scr(j, t)\) , which quantify the relevance of each subprofile in a specific journal j . Regarding our selection strategy, tailored to a single journal focus, follows a more restrictive approach by limiting the number of chosen topics. Particularly, we base this selection on the Dice measure, which, in practical terms, results in the inclusion of few topics, typically one or two. More formally, if the topics related to the candidate journal j in the output ranking are \(t_h, \ldots , t_k\) we use the TopicSelection ( \([w_{t_{h}j}, \ldots , w_{t_{k}j}]\) , Dice).

EE4 and EE5 Explanation text: “ Your research is linked to Developmental Medicine and Child Neurology journal through topic 56 . The terms highlighted (red/bold face) in your submission are the ones that played a significant role in determining this association. The word cloud showcases the most frequently used terms in those papers that, under this topic, have been published in this journal ”

4.2.3 EE5: related terms

The final explanation element associated with a given journal also relates to the set of selected topics, but in this case, we aim to concentrate our attention on those significant terms in the target article that contribute the most to the relevance of each selected subprofile. In order to do so (see Fig.  5 , right-hand side), the title and the abstract of the target article are displayed with a selection of terms marked in color in order to illustrate the coincidences. In this way, the user can see which article terms are mainly responsible for the journal recommendation and decide whether the system is focusing on the most important ideas or rather on less important ones. This information can help the user decide whether to accept the recommendation or not.

In order to accomplish this, we first isolate the contribution of each term to the subprofile score scr ( j ,  t ) of the selected topic t for the given journal j , whereby terms that do not match the subprofile have a zero contribution. Following de Campos et al. ( 2018 ), we then sort the terms in decreasing order of their contribution and select the ones that achieve at least 90% of the full contribution. We use the cosine measure to compute the similarity between the contributions of both the set of selected terms and the full set of terms. Finally, these selected terms are highlighted in a different color (red) so that they stand out. Additionally, in order to focus the user’s attention on the most significant phrase in the paper, i.e., the one that most contributes to the final score, we also decided to highlight this phrase (in bold).

Continuing with our ongoing example, it is worth noting that when researchers consider both EE4 and EE5, they may decide against submitting their paper to Developmental Medicine and Child Neurology journal. This decision can be driven by the observation that the journal’s scope, under Topic 56, appears to be only loosely aligned to the content of this particular research (mainly generic terms related to children, parents, infants, etc. are highlighted and by themselves can explain the 90% of the resulting score). Making such decision is crucial to prevent publishing the paper in an unsuitable journal which could potentially restrict the visibility and impact of our research.

In Li et al. ( 2021 ), in the context of an RS service, Li et al. detect contexts (places, dates, companion, etc.) and contextual features from reviews of these services (hotels, attractions, venues, etc.) and highlight in the text reviews those relevant for the users according to their profiles. According to their experimental results, this highlighting is very useful for users because it provides a personalized explanation and allows them to identify relevant features of the explanation. The underlying idea of explanation element 5 is similar, but in this case, we only work with plain words and do not refer to any kind of underlying concept or feature.

4.3 Categorizing the explanation elements

Finally, once we have described different explanation elements proposed in this paper, we categorize them according to the different types of explanations and RSs described in Sect.  2 .

Our proposed explanation is basically a model-intrinsic approach since the explanation comes from the recommender model.

In terms of the explanation resources, which are classified as human, item, or feature style, we can conclude that our approach mainly corresponds to the last type since we consider keywords from the article texts and these are used as a kind of journal feature. Since the underlying RS is content based, the recommendation is made based on the similarity between the target article and the journal profiles. However, explanation EE3, which consists of the list of similar articles published in the recommended journal, could be considered as a type of item style explanation.

Our proposal is distinctly local, as it provides an individualized explanation for each recommended item. Additionally, it is non-personalized, given that, despite adapting results to the target paper, we do not consider any information (such as interests or preferences) pertaining to the individual seeking the recommendation. The only information supplied by the user is the title and abstract of the target article. Since the recommendation and explanation are not integrated into our recommender model and the explanations are attached to the recommended items, it is not a recommendation-by-explanation one.

Focusing on the explanation content, we could say that our approach combines the user input (the text provided by the user), the information obtained in the inference process (the recommendation itself), and feature-based information, since keywords are also used to perform the explanation. Moreover, in our case, the explanation style is a combination of content-based and case-based information, named accordingly in Tintarev and Masthoff ( 2015 ), or item style explanation as in Papadimitriou et al. ( 2012 ). The reason for the first is obvious since we are working with a content-based RS, and for the second since similar items to the target are presented for a given recommended journal.

Finally, we would like to emphasize that our system consistently provides accurate explanations for all generated recommendations, as these explanations rely directly on the internal information managed by the RS. The persuasiveness of these explanations is a matter that a user study could unveil. In any case, this may depend on the inherent quality of the recommendations. If the recommendations are not appropriate, it is conceivable that the explanations may lack conviction, for example, featuring topics or highlighting terms that are not central to the target article. However, this could be construed positively, offering users reasons to reject such recommendations.

5 Design of the user study

5.1 objective.

The main objective of the user study described in this section is to determine whether the previously presented designed explanations are deemed to be suitable by biomedical experts and help to understand and accept the RS recommendations.

In line with the explanation benefits outlined by Tintarev and Masthoff Tintarev and Masthoff ( 2007 ), in this study we focus on transparency, satisfaction, trust and scrutability of the explanations, with the further inclusion of quality and novelty of the recommendations. With these aims in mind, we designed the questionnaires to gather user feedback after interaction with the XRS. We do not consider effectiveness, persuasiveness and efficiency since our XRS is not an online, real system enabling actions to be performed once the explanations have been received.

5.2 Dataset

The test collection used in the experimentation is called PMSC-UGR (Albusac et al. 2018 ) and has been created by the authors from PubMed and Scopus. It contains the title, abstract, keywords, citations, and authors of papers published between the years 2007 and 2016 from 1002 different journals. The articles published in the first nine years (a total of 276,679 papers) were used to build the journal subprofiles and feed the RS. The remaining articles from the year 2016 (a total of 32,864) configure the test set from where the articles for showing the recommendations and their explanations were selected in the user study.

5.3 Participants

In order to perform a user study of our recommender system and its explanation facilities, we recruited a number of researchers (mostly university professors) who were responsible for a large number of biomedical publications. Of these, seventeen (10 women and 7 men) completed the user study and evaluated a total of 68 submission recommendations, each one implying the analysis of the 10 suggested journals, which have therefore been considered valid for the result analysis. The distribution according to their expertise is as follows: medical doctors (6), nurse (1), biostaticians (one a researcher from a public hospital) (4), librarian from an hospital library (1), biochemists (2), and computer scientists with expertise in bio-informatics (3).

5.4 Protocol

The objective of the user study had two folds: to assess the quality of both the explanations provided for the recommendations and the recommendations themselves. With this objective in mind, researchers were invited to imagine the following scenario: They had recently written an article and were currently deliberating on the choice of a journal by employing the RS. They were instructed to make the ultimate decision by not solely relying on the suggested journal list but also by scrutinizing the explanations accompanying each recommendation.

To simulate this scenario, the participants were instructed to select some articles from a pool of already published ones. Footnote 9 Below, we detail how this pool of articles has been created and provide the motivations behind it. Now, our focus shifts to detailing how participants reviewed the recommended journals and their respective explanations. For the sake of simplicity, we have omitted some details of the protocol; however, a comprehensive description of the protocol provided to the users for conducting the evaluation is presented in “Appendix Appendix A: .”

Specifically, upon selecting a target article (by clicking a magnifying glass), users were presented with a ranking of 10 journals recommended by the system for that particular article. At this stage, they are not observing the explanation but only a recommendation, the original output for the RS, so the participants can evaluate the list of suggested journals without any explanation. After clicking in the explanation link, the same ranking is offered completed with their corresponding confidence percentages (EE1) and several word clouds of the most related topics (EE2). The user then might click, in turn, on the explanation for each suggested journal to obtain its specific explanations (EE3, EE4, and EE5).

After observing all the explanation for such target article, the participant is requested to complete a specific questionnaire (therefore, there are as many filled questionnaires as target articles evaluated). Once all the selected articles had been analyzed, participants were instructed to finalize the process by completing a global questionnaire at the end.

The questions from these two questionnaires are shown in Tables 1 and 2 . Except for a few questions where an item was selected, all the questions involved scaling responses on a 7-point Likert scale (1—strongly disagree, 2—disagree, 3—somewhat disagree, 4—neither agree nor disagree, 5—somewhat agree, 6—agree, 7—strongly agree). In addition, the global questionnaire contained a question for the users to freely comment on the evaluation.

According to the selected aims of this user study which were outlined at the beginning of Sect.  5.1 , Table 3 provides details about which questions deal with each purpose.

5.4.1 Creating the pool of articles

In addition to the global objective of determining the quality of the explanations, we want to evaluate whether the perceived quality depends on the quality of the recommendations. With this aim, the articles subject to evaluation by participants in the study were chosen from the test set and categorized based on the system’s objective performance. Footnote 10 The classification comprises three categories: Category A includes articles for which the first journal recommended by the system aligns with the actual journal of publication; Category B encompasses articles where the actual journal is among the ten recommended by the system; and Category C consists of articles where the actual journal does not match any of the recommended journals. Importantly, participants were deliberately kept unaware of the actual journal of publication to prevent bias in their assessments. This categorization of articles can be useful for investigating potential correlations between user evaluations of explanations and the system’s ability to accurately identify the journal of publication. The objective is to discern whether such correlations exist within the evaluated articles.

Also, we want to evaluate whether the opinions of the participants are influenced by their expertise in the target research field. To fulfill this purpose, we introduced a fourth Category (D), comprising articles authored by the users participating in the system evaluation, which were directly provided by them. For these articles, the actual journal where the paper was published has been excluded from the lists of recommended journals to mitigate potential bias in users’ opinions stemming from its position in the ranking. This fact was explicitly communicated to the participants. The inclusion of Category (D) is justified by our assumption that assessing explanations for papers they are intimately familiar with would result in higher-quality evaluations, given their in-depth knowledge of the content and topics covered.

The distribution of articles in the pool by category is as follows: A(5), B(14), C(9), D(23). Although these categories were not disclosed to the participants, they were directed to guarantee that each participant selects at least one article from each category. However, responses were not received for all the selected papers. Finally, a total of 68 recommendations and explanations were evaluated by the 17 experts participating in the evaluation. The number of articles evaluated within Categories A, B, C and D is 15, 16, 14 and 23, respectively.

5.5 Results

In this section, the results of the evaluation are presented as well as the insights from their analysis.

Figures 6 and 7 show the charts with the answers to the global and specific questionnaires, respectively. In these figures, the stacked bar charts display the distribution (proportion) of user responses, color-coded to represent different answer categories. In order to facilitate interpretation, each bar is aligned around the neutral answer (neither agree nor disagree) which is marked by a vertical line. This alignment serves as a reference point, allowing for easy identification of negative responses (ranging from strongly disagree to somewhat disagree) and positive responses (ranging from somewhat agree to strongly agree).

Results of the global questionnaire

Results of the specific questionnaire

Table 4 shows the averages and standard deviations of the scores for the questions in the specific questionnaire. Table 5 displays the same information for the questions in the global questionnaire with graded responses.

We can see in these tables that, on average, the evaluation of the explanation capabilities of the system (as well as its recommendations) is positive: the average scores are always between 5 and 6 (between somewhat agree and agree). In particular, the questions which scored close to or greater than 6 are: SQ2, GQ3, GQ7 and GQ8 (relating to transparency); SQ4 and GQ4 (relating to scrutability); GQ5 (relating to trust); and GQ12 and GQ14 (relating to satisfaction). The lowest scored question is GQ2 (5.12), relating to the novelty of the recommendations, although this could be explained by the fact that our evaluators are all experienced researchers in the biomedical field.

In Tables 6 and 7 , we also display, for the specific and global questionnaires, respectively, the number of positive, negative, and neutral answers for each question, where we consider a score of 5, 6, or 7 to be positive, a score of 1, 2, or 3 to be negative, and a score of 4 to be neutral. From these data, using a one-sample test of proportions for each question, we have found that the proportion of positive answers is significantly greater than the proportion of negative answers in every case (with p-values that are mostly very low). This confirms that our system is positively evaluated by the users.

Focusing on the specific questionnaire outlined in Table 6 , it is evident the positive impact of explanations when comparing the responses to SQ1 with those of other questions. Thus, while the recommendations, without the presence of explanations, were initially received positively (51), the counts for other specific questions, which consider the impact of the explanation, increased. Conversely, the opposite trend is observed for instances of negative responses to SQ1 (17). These findings confirm the positive role that explanations play in users’ perception of the RS.

Digging deeper, our goal is to investigate whether users’ subjective opinions on the suggested journals have an impact on their perception of the utility of the explanation. For this purpose, in Table 8 , we also present how users perceived the explanations in two distinct situations: one where the user does not like the suggestions (articles with a negative value for SQ1) and another where the user agrees with them (articles with a positive value for SQ1).

First, let us focus on negative SQ1 answers: Without explanation, the participants gave an average rating of 2.18. However, after evaluating the explanations, their perception of how the RS works (SQ5) increased significantly to an average value of 3.47. Footnote 11 Furthermore, it is noteworthy to mention that the Pearson correlation score between SQ1 and SQ5 is \(-\) 0.673. This score indicates that as SQ1 decreases, there is a greater improvement in user perception after the explanation. This shift can be attributed to the fact that explanations aid in comprehending the RS output (SQ2, averaging 4.29) and considering the sense behind some of the recommendations (SQ3, averaging 3.58). Notably, these explanations prove valuable in their decision-making process (SQ4, averaging 4.76). These are all observations indicating the usefulness of explanations for users.

Conversely, when participants agreed with the provided recommendations (positive SQ1 answers), despite finding the explanations helpful (with an average rating greater than 6 in all cases), there was a marginal decline in their overall perception. The post-explanation SQ5 score equaled 6.07, in contrast to the pre-explanation score of SQ1, which was 6.31, Footnote 12 being in this case positively correlated (Pearson correlation between SQ1 and SQ5 equals to 0.72). One potential interpretation of this scenario is that participants might utilize explanations to deduce that some of the recommended journals lack coherence, leading to a decline in SQ3 compared to SQ1.

These findings suggest that when users agree with the recommendations, the subsequent explanations may appear irrelevant. On the other hand, when users do not like the recommendation, the explanations prove to be beneficial. Similar outcomes have been observed in a different domain as is the design of explanations in recruitment decision support systems (Shulner-Tal et al. 2022 ). Therefore, we want to highlight the contribution of the explanation in those situations where the recommended journals do not align with the user’s criteria, which can be perceived as a weakness for a RS. This perception of inefficacy can lead to decreased trust in the system, dissatisfaction with the user experience, and a reluctance to rely on the recommendations offered by the RS. The results obtained clearly indicate that explanations enhance user satisfaction, helping in their decision-making process. This assertion is corroborated by responses to GQ1, which participants answered after evaluating recommendations for all their selected papers. In this instance, 88.23% of participants found the recommendations to be reasonable, in contrast with the 75% positive responses obtained for SQ1.

In the preceding discussion, we examined the acquired results in relation to subjective user opinions. Now, our attention shifts to the objective quality of the system measured by its capability of suggesting the actual journal. It is worth noting that there may be additional relevant journals, different from the actual journal, where the paper could be published. Footnote 13 However, valuable insights can be gleaned from this analysis. Specifically, in Table 9 , we break down the average scores for the questions in the specific questionnaire into the four categories of articles considered. It appears that there is a clear positive correlation between the opinions of the users concerning the quality of the system (recommendations and explanations) and the objective results obtained. The trends suggest that the better the system objectively performs for an article (Categories C, B, A), the higher users score subjective system performance for this article. For Category D, we obtained intermediate results (mainly between the results for Categories A and B).

In terms of the questions which involved selecting an item and which mainly individually analyzed various explanation elements, Table 10 provides information about which explanation elements are the easiest and which are the most difficult to understand and also which are the most helpful. We can also see that most participants had no previous experience with this type of recommender system.

There are three explanation elements (EE1, EE3, and EE5) that stand out in terms of understandability and usefulness. The easiest to understand is the list of related articles, followed by the numerical scores and the highlighted terms in the title and abstract. The same EEs are also considered the most helpful, reversing in this case the order of scores and related articles. These results seem to tally with the findings of previous work (Carenini and Moore 2000 ; Ehsan et al. 2019 ; Shmaryahu et al. 2020 ) which concur that the preference is for short, easy-to-understand explanations. On the other hand, the users considered the two types of word cloud to be less useful (probably because they have not been understood properly, especially EE4). In this sense, it seems that users need to become familiar with the provided explanation elements to minimize their cognitive effort, aligning with the results in Gedikli et al. ( 2014 ).

In order to conclude this section, we examine different explanation objectives as outlined in Table 3 . Generally speaking, users have reported a positive perception of the system’s quality (GQ1, SQ1). This tallies well with the findings of our offline evaluation (de Campos et al. 2022 ), where the system successfully included the correct journal among the top-10 recommended journals in 71.21% of the cases. Additionally, the users have reported satisfaction with the given explanation (GQ12, GQ14), with an average rating of 6.12. It is of note that this satisfaction is particularly strong when the top-ranked suggestion aligns with the “correct” journal (see Table 9 , Category A, SQ5 column), leading to easily understandable explanations. However, when a recommendation might not tally with the user’s opinions (see Table 9 , Category C, SQ5 column), understanding the explanation becomes more challenging.

The explanations clearly enhance the transparency of the recommendations with an average rating of 5.9 (GQ3, GQ7, GQ8). This perception of transparency seems to be closely tied to the quality of the recommendations themselves (see Table 9 , Category C, SQ2 column, and also GQ6 in Table 5 ). Moreover, these explanations help users reason about the correctness of the recommendation system and enable them to make decisions, as reflected by an average rating of 5.85 for scrutability (SQ4, GQ4). More significantly, when users need to make decisions for their own papers (see Table 9 , Category D, SQ4), the importance of explanations becomes even more pronounced. This scenario represents the real-world use of the RS and emphasizes the crucial role that explanations play since users rely heavily on them to guide their decision-making regarding the choice of journal for submission.

Based on these findings, we can conclude that providing explanations helps users to trust (GQ5) the decisions made by the RS and increases their confidence in its reliability. However, as expected, the quality of the recommendations plays a crucial role in shaping this trust and confidence.

Limitations of this study A possible limitation of this study is the relatively small number of users involved in the evaluation of our explanation proposal. It is important to note that the individuals conducting the study must possess a comprehensive understanding of the subject matter and be familiar with the journals suitable for the publication of the work. The problem is that in the biomedical context where this study was conducted, it is very difficult to find available senior researchers who could serve as volunteers. Nevertheless, despite this drawback, we believe that it is possible to conclude that the explanation approach presented in this paper has been widely accepted as a useful tool for making users aware of the recommendation reasons.

A second limitation arises as a consequence of the aforementioned one, given that the final number of evaluated papers is relatively low. Although we contend that the quantity of evaluated papers is adequate for deriving meaningful results, a larger number of evaluated papers would have enabled more robust and conclusive findings. However, the strength of the obtained conclusions is reinforced by the fact that all participants are senior researchers.

Lastly, during the evaluation, we presented the explanation elements to users as if they were interacting with an operational recommender system. We organized the explanations into two stages (EE1–EE2 initially, followed by EE3–EE4–EE5), where each group of explanation elements could be considered a compact explanation. Within each group, the explanation elements could be displayed in a different order to participants, aiming to avoid potential biases or counterbalancing issues influenced by the presentation order of the explanation elements.

6 Conclusions and future works

In this paper, we have presented a novel explanation approach for complementing the output of a journal CBRS and describing the reasons why the RS suggested such journals for publishing the paper provided by the user.

The elements of the explanation (tag clouds, RS scores, similar papers, abstract highlighting), although not new in themselves, are totally adapted to the current problem and to the RS model, offering an example of a model-intrinsic explanation and combining global and local information extracted from the model.

In order to validate the explanation approach in terms of a number of aims (transparency, satisfaction, trust and scrutability), we designed and implemented a user study in the biomedical domain, where the users, after interacting with the explainable RS, had to complete questionnaires to record their opinions about the process.

From the questionnaire results, we might conclude that, overall, the explanation provided proved useful for understanding why the RS suggested a particular journal, not only as separate elements but as a whole, offering a combined explanation which was accepted by most users involved in the study. This finding, although observed in other contexts, is also documented in a wide range of publications, such as Gedikli et al. ( 2014 ), Tintarev and Masthoff ( 2015 ), Millecamp et al. ( 2019 ), Hernandez-Bocanegra et al. ( 2020 ) and Zhang and Chen ( 2020 ). Also, the study underscores the significant contribution of explanations in scenarios where the recommended journals do not align with the user’s criteria. Similar patterns have been identified in a distinct domain (Shulner-Tal et al. 2022 ).

In addition, we found a positive correlation between the objective performance of the RS and user satisfaction.

Regarding the EEs, it seems that all of them (with the exception of the tag clouds) are easy to understand and offer an added value for interpreting the RS output. However, the two EEs based on tag clouds are more difficult to understand and must therefore be redefined in further research.

In terms of future work, following the suggestions of various users in the observation fields of the questionnaires, we aim to improve the interpretability of the tag clouds as it was sometimes difficult to understand the general concepts or topics that these intended to express. One solution could be to complement the use of tag clouds with a natural language description generated with a large language model to describe the topics integrated in them. Additionally, a number of users claimed that the terms contained were too general and this could in turn lead to confusion and result in an unclear or not useful tag cloud interpretation. We would consequently redefine the selection policy of the tag clouds in an attempt to find the most expressive ones.

Similarly, several users proposed enhancing the explanation by incorporating specific features of the recommended journals, such as the impact factor or other bibliometric indicators. This addition aims to enrich the explanation with supplementary information, potentially improving the decision-making process.

In this context, integrating the journal impact factor (or any other relevant measure) could be accomplished through an online reranking of the initial recommendations, considering the scientometric index of the journals. The decision to include its effect could be delegated to the users, who could utilize a slider, for instance, to adjust the balance between the two extremes (ranging from 100% content to 100% impact factor). Other bibliometric indicators, such as quartiles, can be easily integrated into the recommendation ranking by grouping journals accordingly. The same approach could be applied to the type of journal access; results could be organized by open access and others.

In a previous work, de Campos et al. ( 2022 ), the author’s previous publications were also incorporated into the RS model as an additional factor which also impacts the journal recommendation process. Another interesting line of future research, therefore, would be to incorporate this into the explanation process, with the fact of having previously published in the recommended journals being another key element for explaining them.

Meanwhile, this explanation scheme will be exported out of the biomedical field to cover any type of journal domain and this is a straightforward process. In this way, this kind of explanation, which has proved useful in this domain, would be successfully exported to others, thereby widening the range of researchers who could benefit from the explanation facilities provided.

Finally, we plan to adapt the recommendation system to the problem of academic expert finding (for example, given a scientific text, to recommend researchers whose expertise would be useful for collaborations in the topics covered by the provided text) and to modify the explanation approach to deal with the singularities of this problem.

alternatively, Latent Dirichlet Allocation could also be used, as seen in de Campos et al. ( 2023 ).

with the parameter K denoting the number of clusters considered.

https://lucene.apache.org/ .

h is a parameter, currently fixed to 40.

n is another parameter, presently fixed to 10.

although only 5 terms are displayed in Fig.  3 , this is done to reduce its size.

This is because many subprofiles exhibit some degree of similarity to the query which is composed by the title, abstract, and keywords of the target paper.

In this case, it was applied to select the most representative topics that can be associated with a document which has been characterized by a probability distribution over the entire set of topics obtained from an LDA algorithm.

All of them with a more recent publication date than the articles used to build the system.

This information was intentionally withheld from the users.

Results of the paired-t test indicated that there is a significant medium difference between before and after explanations, \(p = 0.031\) .

Results of the paired-t test indicated that there is a significant small difference between before and after explanations, \(p = 0.009\) .

This can be reflected by the positive outcomes achieved for category C, where users still find value in the list of recommended journals even when the system is unable to retrieve the actual journal.

Afchar, D., Melchiorre, A., Schedl, M., Hennequin, R., Epure, E., Moussallam, M.: Explainability in music recommender systems. AI Mag. 43 , 190–208 (2022)

Google Scholar  

Albusac, C., de Campos, L.M., Fernández-Luna, J.M., Huete, J.F.: PMSC-UGR: A test collection for expert recommendation based on PubMed and Scopus. In: Advances in Artificial Intelligence. CAEPIA 2018, LNAI 11160, pp. 34–43 (2018)

Aletras, N., Baldwin, T., Lau, J.H., Stevenson, M.: Evaluating topic representations for exploring document collections. J. Am. Soc. Inf. Sci. 68 , 154–167 (2017)

Barredo-Arrieta, A., Díaz-Rodríguez, N., Del Ser, J., Bennetot, A., Tabik, S., Barbado, A., Garcia, S., Gil-Lopez, S., Molina, D., Benjamins, R., Chatila, R., Herrera, F.: Explainable artificial intelligence (XAI): concepts, taxonomies, opportunities and challenges toward responsible AI. Inf. Fusion 58 , 82–115 (2020)

Article   Google Scholar  

Bilgic, M., Mooney, R.: Explaining recommendations: Satisfaction vs. promotion. In: Proceedings of the Beyond Personalization Workshop in Conjunction with International Conference on Intelligent User Interfaces (IUI’05) (2015)

Cardoso, B., Sedrakyan, G., Gutiérrez, F., Parra, D., Brusilovsky, P., Verbert, K.: IntersectionExplorer, a multi-perspective approach for exploring recommendations. Int. J. Hum. Comput. Stud. 121 , 73–92 (2019)

Carenini, G., Moore, J.: An empirical study of the influence of argument conciseness on argument effectiveness. In: Proceedings of the 38th Annual Meeting on Association for Computational Linguistics, pp. 150–157 (2000)

Chen, W., Hsu, W., Lee, M.L.: Tagcloud-based explanation with feedback for recommender systems. In: Proceedings of SIGIR Conference (2013)

Chi, J., Ouyang, J., Li, C., Dong, X., Li, X., Wang, X.: Topic representation: finding more representative words in topic models. Pattern Recogn. Lett. 123 , 53–60 (2019)

Daher, J.B., Brun, A., Boyer, A.: A review on explanations in recommender systems. Hal Open Science. https://hal.science/hal-01836639 (2017)

de Campos, L.M., Fernández-Luna, J.M., Huete, J.F.: Committee-based profiles for politician finding. Int. J. Uncertain. Fuzziness Knowl. Based Syst. 25 (Suppl. 2), 21–36 (2017)

de Campos, L.M., Fernández-Luna, J.M., Huete, J.F.: On the selection of the correct number of terms for profile construction: theoretical and empirical analysis. Inf. Sci. 430–431 , 142–162 (2018)

de Campos, L.M., Fernández-Luna, J.M., Huete, J.F., Redondo-Expósito, L.: LDA-based term profiles for expert finding in a political setting. J. Intell. Inf. Syst. 56 (3), 529–559 (2021)

de Campos, L.M., Fernández-Luna, J.M., Huete, J.F.: Publication venue recommendation using profiles based on clustering. IEEE Access 10 , 106886–106896 (2022)

de Campos, L.M., Fernández-Luna, J.M., Huete, J.F.: Use of topical and temporal profiles and their hybridisation for content-based recommendation. User Model. User-Adap. Interact. 33 (4), 911–937 (2023)

Ehsan, U., Tambwekar, P., Chan, L., Harrison, B., Riedl, M.O.: Automated rationale generation: a technique for explainable AI and its effects on human perceptions. In: Proceedings of the 24th International Conference on Intelligent User Interfaces, pp. 263–274 (2019)

Felfernig, A., Tintarev, N., Tran, T., Stettinger, M.: Designing explanations for group recommender systems. arXiv:2102.12413 (2021)

Ferwerda, B., Swelsen, K., Yang, E.: Explaining content-based recommendations. https://www.bruceferwerda.com/docs/Ferwerda_DigitalTVGuide.pdf (2018)

Gedikli, F., Ge, M., Jannach, D.: Understanding recommendations by reading the clouds. In: Lecture Notes in Business Information Processing, vol. 85. Springer (2011)

Gedikli, F., Jannach, D., Ge, M.: How should I explain? A comparison of different explanation types for recommender systems. Int. J. Hum. Comput. Stud. 72 , 367–382 (2014)

Hernandez-Bocanegra, D.C., Ziegler, J.: Explaining review-based recommendations: effects of profile transparency, presentation style and user characteristics. i-com 19 (3), 181–200 (2020)

Hernandez-Bocanegra, D.C., Donkers, T., Ziegler, J.: Effects of argumentative explanation types on the perception of review-based recommendations. In: Adjunct Proceedings of the 28th ACM Conference on User Modeling, Adaptation and Personalization (UMAP ’20 Adjunct) (2020)

Iferroudjene, M., Lonjarret, C., Robardet, C., Plantevit, M., Atzmueller, M.: Methods for explaining Top-N recommendations through subgroup discovery. Data Min. Knowl. Disc. 37 , 833–872 (2023)

Article   MathSciNet   Google Scholar  

Jannach, D., Jugovac, M., Nunes, I.: Explanations and user control in recommender systems. In: Personalized Human–Computer Interaction, pp. 133–158. De Gruyter Oldenbourg (2019)

Jesus, S., Belém, C., Balayan, V., Bento, J., Saleiro, P., Bizarro, P., Pedro, Gama, J.: How can I choose an explainer? An application-grounded evaluation of post-hoc explanations. In: Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency, pp. 805-815 (2021)

Li, L., Chen, L., Dong, R.: CAESAR: context-aware explanation based on supervised attention for service recommendations. J. Intell. Inf. Syst. 57 , 147–170 (2021)

Li, B., Qi, P., Liu, B., Di, S., Liu, J., Pei, J., Yi, J., Zhou, B.: Trustworthy AI: from principles to practices. ACM Comput. Surv. 55 (9), 177 (2023)

Louki, P., Schaffer, J., Pjara, J., O’Donovan, J., Getoor, L.: Generating and understanding personalized explanations in hybrid recommender systems. ACM Trans. Interact. Intell. Syst. 10 (4), 31 (2020)

Lully, V., Laublet, P., Stankovic, M., Radulovic, F.: Enhancing explanations in recommender systems with knowledge graphs. Procedia Comput. Sci. 137 , 211–222 (2018)

Millecamp, M., Naveed, S., Verbert, K., Ziegler, J.: To explain or not to explain: the effects of personal characteristics when explaining feature-based recommendations in different domains. In: IntRS ’19: Joint Workshop on Interfaces and Human Decision Making for Recommender Systems (2019)

Minh, D., Wang, H.X., Li, Y.F., Nguyen, T.N.: Explainable artificial intelligence: a comprehensive review. Artif. Intell. Rev. 55 , 3503–3568 (2022)

Nunes, I., Jannach, D.: A systematic review and taxonomy of explanation in decision support and recommender systems. User Model. User-Adap. Interact. 27 (3–5), 393–444 (2017)

Papadimitriou, A., Symeonidis, P., Manolopoulos, Y.: A generalized taxonomy of explanation styles for traditional and social recommender systems. Data Min. Knowl. Disc. 24 , 555–583 (2012)

Pavitha, N., Pungliya, V., Raut, A., Bhonsle, R., Purohit, A., Patel, A., Shashidhar, R.: Movie recommendation and sentiment analysis using machine learning. Glob. Transit. Proc. 3 (1), 279–284 (2022)

Pérez-Núñez, P., Díez, J., Bahamonde, A., Luaces, O.: Text-based recommender system with explanatory capabilities. ResearchSquare. https://doi.org/10.21203/rs.3.rs-1536768/v2 (2022)

Polleti, G.P., Cozman, F.G.: Explaining content-based recommendations with topic models. In: 8th Brazilian Conference on Intelligent Systems, pp. 800–805 (2019)

Radensky, M., Downey, D., Lo, K., Popovic, Z., Weld, D.: Exploring the role of local and global explanations in recommender systems. In: CHI’s Extended Abstract. arXiv:2109.13301 (2022)

Rana, A., D’Addio, R.M., Manzato, M.G., Bridge, D.: Extended recommendation-by-explanation. User Model. User-Adap. Interact. 32 , 91–131 (2022)

Sato, M., Nagatani, K., Sonoda, T., Zhang, Q., Ohkuma, T.: Context style explanation for recommender systems. J. Inf. Process. 27 , 720–729 (2019)

Shmaryahu, D., Shami, G., Shapira, B.: Post-hoc explanations for complex model recommendations using simple methods. In: Joint Workshop on Interfaces and Human Decision Making for Recommender Systems. CEURS, 2682:paper3 (2020)

Shulner-Tal, A., Kuflik, T., Kliger, D.: Fairness, explainability and in-between: understanding the impact of different explanation methods on non-expert users’ perceptions of fairness toward an algorithmic system. Ethics Inf. Technol. 24 , 2 (2022)

Sullivan, E., Bountouridis, D., Harambam, J., Najafian, S., Loecherbach, F., Makhortykh, M., Kelen, D., Wilkinson, D., Graus, D., Tintarev, N.: Reading news with a purpose: explaining user profiles for self-actualization. In: Proceedings of 27th Conference on User Modeling, Adaptation and Personalization Adjunct (UMAP’ 19 Adjunct) (2019)

Tintarev, N., Masthoff, J.: A survey of explanations in recommender systems. In: IEEE 23rd International Conference on Data Engineering Workshop, Istanbul, Turkey, 2007, pp. 801–810 (2007)

Tintarev, N., Masthoff, J.: Evaluating the effectiveness of explanations for recommender systems. User Model. User-Adap. Interact. 22 , 399–439 (2012)

Tintarev, N., Masthoff, J.: Explaining recommendations: design and evaluation. In: Recommender Systems Handbook, chapter 10, pp. 353–382. Springer (2015)

Tsai, C., Brusilovsky, P.: Explaining recommendation in an interactive hybrid social recommender. In: IUI ’19: Proceedings of the 24th International Conference on Intelligent User Interfaces March, pp. 391–396 (2019)

Verbert, K., Parra, D., Brusilovsky, P., Duval, E.: Visualizing recommendations to support exploration, transparency and controllability. In: Proceedings of the 2013 International Conference on Intelligent User Interfaces (2013)

Vig, J., Sen, S., Riedl, J.: Tagsplanations: explaining recommendations using tags. In: Proceedings of Intelligent Users Interface Conference. pp. 47–56 (2009)

Vultureanu-Albisi, A., Badica, C.: A survey on the effects of adding explanation to recommender systems. Concurr. Comput. 34 (20), e6834 (2022)

Yao, Y., Wang, C., Li, H.: Counterfactually evaluating explanations in recommender systems. arXiv:2203.01310 (2022)

Zhang, Y., Chen, X.: Explainable recommendation: a survey and new perspectives. Found. Trends Inf. Retr. 14 (1), 1–101 (2020)

Zhang, Z., Chen, L., Jiang, T., Li, Y., Li, L.: Effects of feature-based explanation and its output modality on user satisfaction with service recommender systems. Front. Big Data 5 , 897381 (2022)

Download references

Acknowledgements

This work was jointly funded by MCIN/ AEI /10.13039/501100011033 under project PID2019-106758GB-C31; the State Research Agency (SRA) and European Regional Development Fund (ERDF) under project PID2022-139293NB-C33, and the Spanish “FEDER/Junta de Andalucía-Consejería de Transformación Económica, Industria, Conocimiento y Universidades’ under Grant A-TIC-146-UGR20, and the European Regional Development Fund (ERDF - FEDER Una manera de hacer Europa).

Funding for open access publishing: Universidad de Granada/CBUA. This work was jointly funded by MCIN/ AEI /10.13039/501100011033 under project PID2019-106758GB-C31 and the Spanish “FEDER/Junta de Andalucía-Consejería de Transformación Económica, Industria, Conocimiento y Universidades” under Grant A-TIC-146-UGR20, and the European Regional Development Fund (ERDF - FEDER Una manera de hacer Europa).

Author information

Luis M. de Campos, Juan M. Fernández-Luna and Juan F. Huete have contributed equally to this work.

Authors and Affiliations

Departamento de Ciencias de la Computación e Inteligencia Artificial, Escuela Técnica Superior de Ingenierías Informática y de Telecomunicación, CITIC-UGR, Universidad de Granada, C/ Periodista Daniel Saucedo Aranda, S/N, 18013, Granada, Spain

Luis M. de Campos, Juan M. Fernández-Luna & Juan F. Huete

You can also search for this author in PubMed   Google Scholar

Contributions

All authors contributed to the study conception and design. Material preparation, programming, data collection, and analysis were performed by the three of us. The first draft of the manuscript was written collaboratively, and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Corresponding author

Correspondence to Juan M. Fernández-Luna .

Ethics declarations

Conflict of interest.

The authors have no relevant financial or non-financial interests to disclose.

Additional information

Publisher's note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

This paper or a similar version is not currently under review by a journal or conference. This paper is void of plagiarism or self-plagiarism as defined by the Committee on Publication Ethics and Springer Guidelines.

Description of the protocol of the user study

Before commencing the evaluation, users were instructed to review the study’s guidelines to familiarize themselves with various EEs, understand their interpretation, grasp the flow of the explanation, and learn how to conduct assessments.

To initiate this process, we provided users with the following introductory background regarding the issue of venue recommendation:

Background Given a paper ready to be submitted to a journal or conference, venue in general, the problem of venue recommendation attempts to automatically identify suitable publication venues for such paper and suggest them to the author. This is a difficult problem for various reasons: firstly, because there is a huge number of possible publication venues, and secondly, because even within a single specific research domain, there are thousands of publications. It is not, therefore, easy for researchers to be aware of every academic venue that would suit their domain of interest. The problem is further exacerbated by the increasing number of papers which contain multidisciplinary research and by the dynamic change in the scope of certain venues. The situation is also more difficult for new inexperienced researchers and for experienced researchers who move to new research areas. In a previous research, we developed a journal recommender system that tries to ease this task by suggesting suitable journals that the author could consider for publication. In this new research, we are conducting a user study to check whether explanation features designed for the journal recommender system are useful for understanding the recommendation and increasing the user’s confidence on it.

The second part is focused on the aims of the evaluation:

Aims At this point, it is important to remark that we are trying to evaluate separately the quality of the recommendations offered by the system (i.e., whether the proposed journals seem you appropriate to publish the given article) and the quality of the explanations offered by the system to justify its recommendations. Our main goal is to evaluate the quality of the explanation facilities of the system, although knowing your opinion about the quality of the recommendations is also important. It should be noticed that it is possible that the system makes correct recommendations and at the same time the explanations of these recommendation are useful, but it can also be the case that the recommendations are okay but the explanations are poor and also that the recommendations are not appropriate but the explanations of these bad recommendations make sense (and even can be useful to decide not to accept the recommendations). Both bad recommendations and bad explanations are also possible. This system, in its current state, represents a first step within the journal selection process by recommending journals that are suitable to publish a given paper based solely on thematic content. In this sense, when you are deciding whether the journal recommendations offered by the system are good or not, the decision should be made based exclusively on the thematic content of the articles published in the recommended journal(s) compared with that of the target article, without taking into account other factors concerning the journals, as for example impact factors, editorial boards or duration and hardness of the evaluation process.

Next, users were provided with a comprehensive description of all the EEs they would encounter in the evaluation. Additionally, instructive examples in the form of graphics, as illustrated in Figs.  2 , 3 , 4 , and 5 from Sect.  4 of the paper, accompanied the descriptions. These examples aimed to illustrate specific elements and guide users on how to interpret them, serving as explanations for the recommended journals. The objective was to acquaint users with the EEs and enhance their understanding of how to interpret them in the context of explanations:

Explanation Elements [EE1] —In the list of recommended journals, attached to each journal there is a percentage that reflects the confidence of the system on the recommended journal, i.e., showing decreasing degrees of matching between the target article and each of the journals, as estimated by the system. These scores explain the ranking of recommended journals provided by the RS as they give information about the certainty of the RS on each recommended journal. [EE2] —An association of the target article to one or several thematic/topical areas which are commonly covered by the recommended journals. These thematic areas are represented in the form of word clouds. The words represented in them are the most important ones for those topics. These word clouds explain the main topics of the recommended journals in relation to the target article and the user could use this information to assess whether the topics of the target article are related to one or several topics extracted from all the journals in the ranking. If not, it could be reasonable to think that the recommendation might not be totally useful as the journals do not publish papers in the target article’s topics. This EE does not depend on a specific recommended journal. [EE3] —A list of the up to three most similar articles published in a selected journal from the ranking (local to a chosen journal). Each similar article has associated a traffic light highlighting the green, orange or red lights according to the degree of similarity with the target article (green, very similar; orange, medium similarity; red, slightly similar). Three very similar papers to the one to be submitted (all in green color) would explain the fact that this journal has already published papers with closely related content so it would be a good option to submit the target article. Otherwise, if they are in red, they are less related and therefore the journal may be less appropriate than others, although it still could be a viable option. [EE4] —One or several word clouds which reflect topics commonly covered by the articles published in a selected journal (local to a chosen journal). If the journal topics shown are similar to the topics covered by the target article, it could be concluded that the target article would fit very well in the scope of the journal. [EE5] —For each of the shown word clouds/thematic areas (EE4), the abstract of the target article with those words in common with this area colored, in order to illustrate the coincidences. The more words colored in red in the abstract more coincidences with the words forming the topic, so more appropriate the journal could be for publishing.

Finally, the users could read a text with the flow of the evaluation:

Evaluation Process To begin with the evaluation, a list of preloaded target articles are shown grouped in four groups. The user is asked to evaluate several of them from each group (at least one from each category). After clicking in the corresponding target article link (magnifying glass), a first page is shown containing the information of the target article (title, abstract, keywords and concepts) and the recommendation in the form of a ranking of 10 journals. This is not part of the explanation, only the recommendation itself. The user should click on a link presented in this page for the explanation. After this action, the global explanations for that target article and their 10 recommended journals, i.e., EE1 and EE2, are shown. In the EE1 ranking, the user may click on the link of each recommended journal for receiving its local explanation, giving way to explanations EE3, EE4 and EE5 (specific of that journal). Once the evaluation of a target article is finished, a questionnaire is requested to be filled, containing specific questions related to this last evaluation (link found at the right top corner of the web page of the explanations). Please, be aware that this questionnaire must be completed for each target article selected. This process could be repeated for each target article as many times as desired by the user. After iterating this process for several target articles, and interacted with the recommendations and their explanations, a final questionnaire with questions related to the whole evaluation process is asked to be filled, which is found at the left main menu. At this point the evaluation is finished.

When engaging with the RS and its explanations, to serve as a reminder and aid in understanding the significance of each EE, accompanying descriptions were provided. These descriptions can be found in the captions of Figs.  2 , 3 , 4 , and 5 , corresponding to Sect.  4 of the paper.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ .

Reprints and permissions

About this article

de Campos, L.M., Fernández-Luna, J.M. & Huete, J.F. An explainable content-based approach for recommender systems: a case study in journal recommendation for paper submission. User Model User-Adap Inter 34 , 1431–1465 (2024). https://doi.org/10.1007/s11257-024-09400-6

Download citation

Received : 27 June 2023

Accepted : 17 May 2024

Published : 06 June 2024

Issue Date : September 2024

DOI : https://doi.org/10.1007/s11257-024-09400-6

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Explainable recommender system
• Publication venue recommendation
• Find a journal
• Publish with us
• Track your research

Top 10 Compelling Content Marketing Case Studies

The World of Marketing is changing every hour! Trends, technologies, and consumer needs are constantly evolving. Staying relevant in such times is extremely important. And what better way to do that, than a super-strong content marketing strategy!

Content Marketing is not just confined to building trust, awareness, or engagement. Today it has become the new normal of the marketing world. Software companies have now positioned themselves as tech-educators through blogs, how-to videos, and informational guides. While global tech giants like Google have come up with free courses and certifications to educate their audiences in various domains from Digital Marketing to Artificial Intelligence. Such content is a great way to remain relevant, inform target audiences and convert potential customers into loyal ones. 

With the help of engaging Content Marketing Strategies, brands can highlight their business’s key services, strengths, USPs and drive brand utility in the daily lives of their customers.

“Content marketing is a strategic marketing approach focused on creating and distributing valuable, relevant, and consistent content to attract and retain a clearly-defined audience — and, ultimately, to drive profitable customer action.” Content Marketing Institute

It means qualitative, consistent, and relevant content drives the purchase decisions of the consumers and helps brands generate leads. Content can be generated in various formats such as blogs, newsletters, surveys, tool-reviews, e-books, quizzes, videos, giveaways, podcasts, or user-generated content.   

Why Content Marketing is Important? 

Here are the top three reasons why you should consider content marketing over any other marketing channel. 

1) For Building Trust and Brand Reputation

When consumers read about your brand, they tend to develop an opinion about you. If the content provided is valuable, educational, and engaging, consumers are more likely to invest in your products or services. Therefore articles or blog posts build trust and a good reputation for your brand in eyes of targeted audiences.

2) Helps Generate leads and Influencing Conversion Rates  

YES! content marketing is effective in generating leads too and also works well in improving website conversion rates. Improved conversion rates lead to a higher probability of increased sales.  

“According to 74% of companies surveyed, content marketing has increased their marketing leads, both in quantity and quality. “(Curata)

“Content marketing provides conversion rates about 6 times higher than other digital marketing methods.”(ABG Essentials)

3) Cost-Effectiveness  

If you aren’t already helping your brand with content marketing, it is high time that your brand starts. Invest time and money into content marketing to see cost-effective and better ROI on your marketing efforts. 

How about we take a look at some successful Content Marketing Case Studies to draw some inspiration. These Content Marketing Case studies will help you analyze the effectiveness of the strategy applied and help create better plans. Need help planning Content Marketing strategies, feel free to connect with us at  [email protected] .

Content Marketing Case Studies

1) company: superdrug.

Superdrug is a health and beauty products retail company based in the UK. It offers a range of products from fragrances, makeup, skincare, toiletries to hair care products and much more. 

Objective: 

The campaign was designed around this one question. – “How do perceptions of beauty vary across the globe?”

• Understanding unrealistic beauty standards and body image roles in different countries. 
• Raising awareness on the issue of body imaging and acknowledging people being held as victims of such issues. 

The Strategy:

Superdrug started a campaign named, “PERCEPTIONS OF PERFECTION” . As part of the campaign strategy, a woman’s image was sent to 18 female graphic designers across different countries. Superdrug asked them to photoshop the image to look attractive to the people of their country. Later, Superdrug created a visual asset compiling the different versions of the original image produced by graphic designers from 18 countries. 

You can check the compilation here- Perceptions of Perfection | Superdrug™

The Results:

• The campaign gained international appreciation and was picked up by 600 publishers including Buzzfeed, Huffington Post, CNN, New York Times, and many more placements. 
• Nearly 1,000,000 social shares.
• More than 700,000 page views on Superdrug’s website. 
• 238% increase in organic traffic search over 16 months period. 
• Mentioned by celebrities like Sofia Vergara and Lil Wayne. 

Key Takeaway: 

• Producing emotionally appealing content is one of the best ways to connect and encourage engagement with your customers. 
• Customers feel that you empathize with them and understand their dynamic perceptions. 
• Moreover, it puts your brand in the limelight for being relevant, acknowledging the challenges faced by customers and helps create better relationships. 

2) Company: Capgemini

Capgemini is a global consulting, technology services, and digital transformation company based in France, offering worldwide enterprise services. 

Challenge- Building Brand Awareness and Reputation Management . 

Despite being capable and working with the world’s top brands including KPMG, Deloitte, and Accenture, the company was falling behind its competitors. They invested large sums of money in paid advertising, appointed professional golfers for branding, and yet faced slow revenue growth with no ROI. 

• Attracting new visitors to the website. 
• Building Brand awareness. 
• Providing knowledge and tech-driven solutions to the target audience. 
• The Strategy: 

The brand created a content marketing strategy plan that revolves around engagement. They converted website content into a storytelling format that would inform and answer the target audience’s questions on the topics like cloud technologies, AI intelligence, Big data, etc… The brand acknowledged the challenges faced by the enterprises and provided them with solutions through their website content. They framed a series of stories, blogs, and informational pieces for their website, educating their audiences about the need for technology driven solutions.  

Source- marketinginsidergroup.com

• In one year’s duration, Capgemini accomplished the goal of attracting 1 million new visitors. 
• Attracted 100,000 new followers for LinkedIn page.
• 1.8 million shares of the content. 
• The strategy was responsible for generating nearly $1 million in sales and $5 million in its second year. And is today generating around  $20 million in sales a year. 

Key Takeaway:

• Information and Problem Solving is the key to attract your customers when talking about new technologies and advancements. 
• Big budget advertisements may not be fruitful if targeted audiences are not well-informed about the product/company. 
• Tech brands and service brands can easily crowdsource informational blogs and articles, share client’s stories, make how-to-videos or create informational e-books. This kind of content becomes more reliable and drives more engagement. 

3) Company: Mercedes Benz

The company needs no introduction. Mercedes-Benz is a German- automotive leader, known for producing luxury vehicles and commercial vehicles. 

• Targeting the younger generation and enhancing brand awareness across social media handles. 
• Generating social media impressions for the new launch- Mercedes CLA. 

Mercedes Benz wanted to reach out to younger audiences. Their strategy concentrated on user-generated content encouraging engagement with the brand. 

Mercedes Benz collaborated with five Instagram photographers and asked them to take photographs of the all-new Mercedes CLA under the “Take the Wheel” challenge. The best part of the campaign was, whosoever gets the maximum likes keeps the car. And hence all five of them worked hard and in turn Mercedes generated some stunning visual assets for their social media handle with the worldwide popularity of the campaign. 

Check out the video! 

• 87,000,000 organic Instagram impressions
• 2,000,000 Instagram likes
• 150 new marketing assets (stunning photos)
• The campaign highlights the importance of choosing the right type of platform and social media influencers for your targeted audiences. 
• Drive more from your contests and generate as much content out of it as possible. 
• And in turn, your brand can generate qualitative and quantitative marketing assets that will keep attracting attention from your target audiences.

4) Company: Colgate Palmolive 

Colgate Palmolive is an American global brand with a history of more than 200 years. The product range includes household, healthcare, personal care, and veterinary products.

Objective:  

• Enhance brand awareness and boost online sales of Palmolive Gift sets. 
• Better Engagement through interactive gamification content. 

Colgate used an interactive content strategy for boosting sales of its Palmolive gift sets during one of its Christmas campaigns. Colgate launched Bubble-popping gamified content combining it with Augmented Reality to engage its customers. To play, users had to present proof of their purchase and then a landing page would lead them to a game. As part of the game, the player has to pop the bubble with his/her nose in the bath like setting. 

Source- contentmarketinginstitute.com

The Result:

• Attracted over 1 million participants within two weeks through Facebook. 
• 10,000+ participants landed on the retailer’s website to shop online. 
• Creative content marketing takes your brand ahead in an overcrowded marketplace. Try the many interactive content marketing formats like games, AR and VR applications, quizzes, and more that are available today. 
• According to Think with Google, “People today have 2X more interactions with brands on mobile than anywhere else.” The statement justifies the increase in the importance of mobile-friendly content marketing strategies.   

5) Company: SEMrush 

SEMrush is a leading competitive research toolkit company that offers its services in PPC, content marketing, SEO, analytics, social media, and a wide range of digital marketing related services. 

• To educate the audiences about the new feature launch. 
• Product awareness and engagement. 

The Strategy :

SEMrush designed an interactive game that educated its target audience about the new additional features on SEMrush’s toolkit. The game was named; “SEMrush Easter Egg Hunt Game ”, it involved the completion of a set of actions within SEMrush’s tool to find 15 easter eggs. And later on, people shared their accomplishments through Twitter. 

Check out the Challenge video here; Virtual SEMrush Egg Hunt Tour

You can check out more information here- SEMrush Easter Egg Hunt: All the secrets revealed!

• Attracted more than 9,300 participants. 
• Generated 8 million Twitter impressions.
• Adding advanced features to your products is easy; educating your targeted audiences about them is tough. Therefore, planning a content marketing strategy for educating your audiences becomes inevitable. 
• Invest in gamified content for brand education. It helps drive brand messages more effectively, attracts more customer eyeballs, and in turn informs them about your offerings.  

6) Company: Swiggy 

Swiggy is India’s leading, tech-driven, on-demand food delivery platform. It brings food from neighborhood restaurants, cafes and eateries directly to your doorstep. It has appointed delivery partners all over India to deliver food through an app-based system. 

• Positioning Swiggy as a compelling choice for delivery partners and encouraging potential delivery partners. 
• Engaging its delivery partners and encouraging them to showcase their creative side.
• A second income generating source for Swiggy’s delivery partners. 

Swiggy Starhunt , the campaign launched by Swiggy acknowledging its delivery partners’ effort. The campaign let it’s delivery partners take a break from their daily work and showcase their talent through Tiktok. (though it has been banned in India now, there are various applications in the market to replace it.)

Swiggy launched four teaser videos to attract customers, which gained remarkable views.

Swiggy Starhunt – Simran

The Star Hunt Challenge included categories like; dancing, singing, musical instrument, and acting. The winners were promised exciting prizes and holiday packages. 

Swiggy used 11 vernacular languages to promote the challenge across pan-India audiences. And collaborated with Tiktok influencers to increase the reach of the challenge. 

The Results:  

• Through Tik Tok influencers, Swiggy garnered an organic reach of 57 Lakhs. 
• 350+ delivery partners participated in the challenge. 
• 1500+ videos were uploaded on Tiktok. 
• 44M+ views on Tik Tok #SwiggyStarHunt. 

Key Takeaway- 

• Producing video content is exciting and an awesome tool to tell stories about your product.
• Cisco projected that more than 80% of all Internet traffic will be video by 2021—which means brands definitely need a video content strategy. (Cisco)  
• Moreover, 87% of consumers want more videos from brands. (Hubspot) . As a brand, you must not overlook the efficiency of rising video platforms like Instagram reels, You Tube Shorts, or Vimeo.  

7) Company: MIVI

MIVI is a homegrown Indian electronics manufacturer company. Its product range includes; headphones, ear pods, speakers, charger cables, and adapters. 

• Product Awareness for new MIVI Duopods M80. 
• Product launch and engagement. 

MIVI made exceptional use of the Instagram platform by launching a story game. The game revolved around completing a mission by carrying all new MIVI Duopods to the safe house. The game was Insta story-based, where players have to tap on the right option that would lead them to the safe house. After completing the mission, users can fill in their information and receive exciting offers, and stand a chance to win MIVI Duopods M80. 

MIVI reached out to various social media influencers like Bhuvan Bam, Mostlysane, Carry Minati, and Mumbaikar Nikhil. They shared the game on their social media handles and encouraged audiences to take up the mission. 

Source- nextbigbrand.in

The Results: 

• More than 3 lakh people played the game in a span of 24 hours. 
• Gained massive social media shares. 
• Interactive content is what drives brand engagement today. Craft interactive content strategies to see your users have fun with the brand.
• Interactive content allows audiences to actively participate rather than passively read, watch or listen to ads. 
• It also enhances the customer experience. The better the customer experience is, the higher are the chances for increased conversion rates.

8) Company: Blendtec

Blendtec is a US-based blender company manufacturing blenders for commercial and residential purposes. It guarantees to offer the safest, highest quality, and speedy blending services. 

• To position Blendtec’s blenders as the world’s most efficient blenders. 

The Strategy:  

Their strategy has been the same for more than a decade. Engaging audiences by blending anything. It includes iPhones, keys, sketch pens, remotes, marbles, or literally anything. The brand keeps posting “Will it blend?” Youtube videos and has been successfully creating viral content for around 13 years. 

Check out Will it blend? Videos here! 

• 700% increase in sales in recent years. 
• More than 850k subscribers on the Youtube channel. 

Make your product your hero in your content marketing strategy. And drive the content around it. Get your audiences hooked to your content and create a permanent brand recall for your product in the mind of the audiences.

9) Company: Airbnb 

Airbnb is an American vacation rental company that connects people with places, stays, and hosts through a website and app system. 

• Increasing organic reach through content that satisfies all travel needs. 
• Encouraging engagement, and awareness about the Neighborhood Guides provided by Airbnb. 

As part of the content creation strategy, they designed a Neighborhood guide for travelers that would help them decide where to stay, what experiences they will gain, and what places their neighborhood would have. The Neighbourhood guide was published on their website and mobile application, which still gets updates from time-to-time by their registered hosts. The guide gives recommendations about food, stays, nearby destinations. A complete guide for every type of traveler!    

You can check their intriguing Neighbourhood Guide here!

• 10,000+ estimated monthly organic traffic. 
• 100+ referring domains. 
• Customize content production focusing on the specific needs of your consumers. This will help your brand increase its credibility and engagement. 
• 78% of CMOs believe custom content is the future of marketing. (DemandMetric) Brands should focus on investing resources in understanding consumer needs and adopting personalized content practices like the website content designed by Airbnb.  

10) Company: Lenovo 

Lenovo is a Chinese multinational technology company offering its products including laptops, ultrabooks, tablets, and other technological devices. 

• To guide customers through the content marketing funnel. Firstly by building awareness about the brand, educating about products, and then converting them into customers. 
• Influencing and educating buyer’s about the technology and helping them with purchase decisions. 

Lenovo created a digital content hub called Tech Revolution. It focused on informing people in the Asia Pacific region about new technologies, trends, and their usages. They built an online website, https://www.techrevolution.asia where they shared stories and informational articles educating their audiences about the products available. Lenovo also collaborated with IT professionals to create informational guides about the continuous evolution of technology.

Checkout Frizbee’s video they created as part of the Tech Revolution!

• 500+ articles with 34 million impressions. 
• 250k+ link clicks
• 60k+ social engagements. 
• 100k+ new visitors. 
• Consumers want to know if they are dealing with the experts in the field. And the best way to assure your audiences is by providing educational content. 
• Educational Content helps your customers make informed purchase decisions and in return builds your brand’s credibility. 
• Digital customers love sharing their reviews and opinions. User-generated information further helps your brand cut costs and gain consumer trust. 
• Consumers are more likely to trust testimonials, expert advice, or other informational guides stating the value of your products. 

Plan your Content Strategy and generate Conversions! 

These are some of the most compelling content marketing strategies by brands across the globe. From emotional appeal to engaging content marketing plans, these are some of the brands that have successfully leveraged the power of Content Marketing. And continue to generate results by creating qualitative, interactive, and relevant content. 

If you are a brand looking for digital marketing agency in Mumbai whose offer content marketing services, reach out to us at  [email protected] . And we will help you create awesome content strategies that convert. Cheers! 

Related Posts

The Benefits of Social Media Contests and Giveaways: How to Boost Engagement and Drive Sales

The Power of Social Media Advertising: Strategies for Boosting Your ROI

Top 8 Social Media Post Ideas for International Women’s Day

Paid social media: Why Use Paid Advertising to Drive Social Media Success?

Leave a reply cancel reply.

Your email address will not be published.Required fields are marked *

Hit enter to search or ESC to close