Recommendation Systems: Using Recommendations for Decision Making
This is part 4 of our series on Recommendation Systems, previous segments are available in our archives.
I’m sure that when you think of recommendations your first thought is of those “People also bought” prompts when buying something online. However, recommendation systems are much more powerful than such a simple application. Why?
Recommendation systems are predicting customer satisfaction.
Instead of thinking as recommendation systems as a suite of recommendation products, think about what we have learned about how they work. The output of a recommendation system is an estimate of how likely a customer is to like a product enough to purchase/consume it. This has a vast number of applications for your business:
- New Product Testing. When developing any new product, you will first test it with a small group of test customers. If that small group is not representative of the interests of your larger customer base, you can get misleading results! A recommendation system can naturally identify how closely the interests of your focus group align with your full population of customers and give you an indication of whether the lessons will scale.
- Customer Segmentation. Recommendation systems will naturally cluster customers based on the kinds of products they like and might like in the future. This can be an important factor in building customer segments that might not be obvious from other features of the customer. Such interest-based clusters are typically the best clusters to target for optimization because you know what is motivating them to use your products.
- New Product Development. If you use a featured recommendation system, as discussed yesterday, it not only tells you about customer satisfaction, but also which features most influence that satisfaction. Those factors can be important inputs to your new product development process as the key factors to focus on to appeal to customers. For example, if price is one of the most important features that predict customer satisfaction, you know you need to focus on price as a core feature of the new product.
There are, of course, many more applications! Recommendation systems are used in a wide variety of applications in business today to do more than just improve products.
Speaking of which, tomorrow we’ll finish our survey of recommendation systems by talking about how they can improve your products.
Quote of the Day: “It’s really hard to design products by focus groups. A lot of times, people don’t know what they want until you show it to them.” – Steve Jobs