Recommendation Systems: A guide that goes beyond traditional e-commerce use cases

A joyful shopper holding a 'Suggested For You' shopping list in a grocery store, symbolizing the personalized touch of recommendation systems

In a marketplace saturated with choices, the ability to connect consumers with the right product or service is paramount.

This is the place where recommendation systems shine, guiding choices with precision and insight. Their prowess is well-known in e-commerce, media streaming, and social networking.

However, the true potential of recommendation systems lies in their application across varied industries, especially for businesses that have invested in rich data ecosystems and robust data warehousing.

But we’re getting ahead of ourselves, let’s start by first clarifying … 

What actually are recommendation systems?

Recommendation systems serve as intelligent filters and matchmakers, analyzing vast datasets to unearth patterns and preferences. By doing so, they can predict and suggest products, services, and actions that resonate with individual users. For businesses, these systems can transform a passive data collection into an active tool for driving engagement, satisfaction, and sales.

The efficacy of these systems is not rooted in the data itself but in the nuanced ways they can interpret and apply it. Companies that have diligently captured and structured their data are now positioned to capitalize on the advanced capabilities of recommendation systems.

What are the types of recommendation systems?

Collaborative Filtering – “Customers who are similar to you, also liked …”

Does this sound familiar? This is the baseline of this recommendation system type. If two users liked, used or purchased the same products, the system may recommend a product liked by one user, to the other user (with the condition that the latter didn’t like/use/purchase it yet). Given that Bob and Daniel both listened to Bruno Mars, Eminem and Ed Sheeran on Spotify, but Bob also recently listened to Maroon 5, the system will most likely suggest Maroon 5 to Daniel as well. 

Content-Based Filtering – “If you liked this item, you might also like …”

Imagine browsing your favorite online bookstore. You’ve purchased several historical novels before. The bookstore’s system takes note. Now, when you visit the homepage, it suggests more books from the historical genre. This is content-based filtering in action. It recommends items similar to what you’ve previously enjoyed, simplifying your search for the next great read. 

Market Basket Analysis: “Customers who bought this item also bought …”

Picture this scenario: shoppers frequently pick up bread and peanut butter on the same trip. The Apriori algorithm catches onto this pattern. It uses this information to recommend peanut butter to anyone with bread in their virtual cart.

This approach is a cornerstone of market basket analysis. It’s invaluable for retailers, aiding in everything from crafting product bundles to optimizing store layouts. Plus, it’s a boon for e-commerce sites, powering those “customers who bought this also bought…” suggestions. The Apriori algorithm may not be a recommendation system on its own, but it’s an essential tool in creating systems that offer smart, related product recommendations.

Hybrid Systems

They combined the features of both collaborative filtering and content-based filtering to provide more accurate and relevant recommendations. They are designed to leverage the strengths and minimize the weaknesses of each individual approach. A good example here is TripAdvisor. They offer personalized travel suggestions by considering the user’s past reviews and searched destinations (content-based) and also by analyzing the preferences and behaviors of similar travelers (collaborative filtering).

Starbucks Christmas Edition Cups exemplifying real life business application of context based recommendation systems

Context-Based Filtering

You would probably be happy with this recommendation from Starbucks during Christmas, but it would be strange during another time of the year. Context-based filtering is a type of recommendation system that takes into account the context in which a user operates or makes decisions, in addition to their preferences and behaviors. Unlike traditional recommendation systems that focus solely on user-item interactions, context-based filtering considers additional factors such as time, location, social setting, or the user’s current activity. It is a more dynamic approach that aims to deliver more relevant and timely recommendations by understanding the circumstances under which users might prefer certain items.

Now, after understanding these systems a little bit, let’s get to the juicy part…

How can we apply them? 

We believe that recommendation systems can be applied in a wide range of industries. We are all familiar with a couple of them that we encounter daily: e-Commerce, Streaming Platforms like Netflix or Spotify or Social Media. But apart from suggesting the next movie to watch or the next song to listen to, these systems can be applied in a couple of other industries that might not cross your mind at first: 

Healthcare and Wellness

Suggesting personalized health plans, diets, or workouts based on a person’s health data, lifestyle, and medical history.Companies like Fitbit and Garmin offer wearable devices that track health metrics such as heart rate, sleep patterns, and physical activity. Their associated apps use this data to recommend personalized health and wellness tips, like targeted exercises, sleep schedules, and stress management techniques.

Banking and Finance

An industry that can be revolutionized by offering personalized financial solutions. By analyzing financial behaviors and goals, these systems can offer tailored credit, investment, and wealth management suggestions. They can enable banks to provide customers with loan and credit card options, aligning with individual risk profiles and financial aspirations. Beyond personalized financial advice, these systems also identify cross-selling opportunities for products like insurance and retirement plans.

For financial and investment platforms, recommendation systems can analyze the behavior and success rate of top investors to suggest stocks, bonds, or funds to their users. If a user has a portfolio with certain characteristics, the system could suggest additional securities that complement their existing holdings and align with their risk tolerance, akin to the portfolio strategies used by successful hedge fund managers.

Professional Networking and Job Matching

On professional networking platforms like LinkedIn, recommendation systems can suggest job openings, professional contacts, or articles to read based on a user’s industry, job history, and interaction patterns. For B2B scenarios, such systems might suggest potential leads or partnerships by analyzing business profiles, mutual connections, and endorsement patterns.

Affiliate and Influencer Marketing

In our work with Madrivo, a renowned performance marketing platform, we developed a sophisticated recommendation system to enhance their affiliate marketing strategies. Madrivo connects brands with customers through various channels, offering a network of over 10,000 publishers and advertisers. Their focus is on creating seamless connections and maximizing revenue through performance-based models.

Madrivo proudly displaying the launch of their new offer recommendation system

Our recommendation system for Madrivo utilizes collaborative filtering to suggest high-potential marketing campaigns to publishers.

It analyzes similarities in revenue patterns and campaign successes across different publishers to provide tailored recommendations. The system selects top-performing campaigns, creates a user-campaign matrix for relationship analysis, and applies various filters to ensure campaign appropriateness.

Notably, it addresses the ‘cold start’ problem for new publishers by recommending the most successful campaigns. This approach enhances campaign targeting, leading to more effective affiliate marketing and increased earnings for publishers. Navigate to our marketing automation case study if you’d like to learn more.

Education and E-Learning

Recommendation systems in e-learning platforms can suggest courses, books, or study materials based on a learner’s past engagement, performance data, and educational goals. For instance, if a user frequently enrolls in courses related to data science, the system might recommend advanced courses in machine learning or statistics, or perhaps suggest peer study groups focusing on similar topics.

Checklist: Is a recommendation system Right for Your Business?

  1. Data Availability: Do you have access to rich and diverse data? A recommendation system thrives on data like customer interactions, transaction histories, user preferences, or any other relevant behavioral data.
  2. Diverse Offerings or Services: Does your business offer a wide range of products, services, or content? If you have a varied inventory or a broad spectrum of services, a recommendation system can help navigate your customers to the most relevant options.
  3. Desire for Personalization: Are you looking to provide personalized experiences or solutions to your customers or users? recommendation systems excel in tailoring suggestions to individual preferences, enhancing user engagement and satisfaction.
  4. Need for Enhanced User Engagement: Are you seeking ways to increase user engagement, retention, or time spent on your platform? recommendation systems can significantly boost these metrics by presenting users with relevant and appealing options.
  5. Complex User Decisions: Does your industry involve complex decision-making processes, such as choosing financial plans, healthcare options, or educational courses? recommendation systems can simplify these processes by guiding users towards optimal choices based on their profiles and needs.
  6. Scalability Goals: Are you aiming to scale your operations? recommendation systems can handle growing data and user bases, adapting to changing dynamics and scaling alongside your business.
  7. Integration Capabilities: Can you integrate a recommendation system with your existing technological infrastructure? Effective implementation often requires the system to work in harmony with your current databases, CRM, or other digital tools.
  8. Analytical Insights Requirement: Do you need deeper insights into customer behavior and preferences? recommendation systems not only provide personalized suggestions but also offer valuable insights based on user data analysis.
  9. Competitive Edge Aspiration: Are you looking to stand out in your market? Implementing a recommendation system can be a significant differentiator, offering a more sophisticated, data-driven approach to customer engagement.
  10. Uncertainty about Applicability: Not sure if a recommendation system fits your specific context or industry? Talk to us. We can help you assess your data and operational needs to determine if a recommendation system is a viable and beneficial solution for your business.


Recommendation systems are powerful tools that transcend traditional application boundaries, offering significant value in a wide range of industries. By considering the above checklist, you can gauge whether these systems align with your business needs and objectives.

If you’re still unsure or need expert guidance, don’t hesitate to reach out to us. We specialize in unlocking the potential of data and can help you explore how a recommendation system can revolutionize your business operations.