Some Recommendations on Recommendation Engines

Use Cases & Projects Lynn Heidmann

Today, recommendation engines are ubiquitous. While some are easy to spot, featured in a nice neat “Recommended for You” section, so many more lurk behind the scenes. In fact, it’s entirely likely that most of the time, you don’t even know you’re being subjected to ML-generated recommendations.

Leo Decaprio recommendation engines

There’s a reason for the prominence of recommendation engines — the amount of data, content, and products available across the web keeps skyrocketing, but people spend less and less time on any individual site. A few years back, Time Magazine found that half of their site’s visitors spent less than 15 seconds on the site. Consumers might spend slightly more time browsing when it comes to e-commerce, but with a growing catalog, there’s no way they’re spending enough time browsing to find everything they might be likely to purchase.

But how do you know if you should build a recommendation engine? When is the right moment, and how can you get started? Allow us to make a recommendation (if you will), and for starters, ask yourself these two questions (and after, download our complete guidebook to building a recommendation engine from the ground up):

1. What Is the End Goal of the Project?

Is the idea to build a recommendation engine to directly increase sales overall? Achieve a higher average basket size? Reduce browsing time and make a purchase happen faster? Reduce the long tail of unconsumed content? Increase user engagement time with your product? It’s important to be clear about the ultimate objective so that both business and data teams building the product have the same goals.

2. Is a Recommendation System Really Necessary?

This is perhaps an obvious question, but since they can be expensive to build and maintain, it’s worth asking. Can the business achieve its end goal by driving discovery via a static set of content instead (like staff/editor picks or most popular content)?


Determine if a recommendation engine is the right means to the end goal for your business.

One of the issues that can arise if these basic questions aren’t well considered is that you may spend time, energy, and resources on building a recommendation engine, and you might even have enough data and good initial results, but the recommendation engine only makes very obvious recommendations. Bottom line: if there isn’t enough of a content long tail or no need for the system, perhaps reconsider the need to build a recommendation engine in the first place.

If You Decide a Recommendation Engine Is Right for Your Business…

There are still plenty of questions, considerations, and challenges ahead. Since there isn’t one single type of recommendation engine, you’ll need to figure out the best strategy (or combination of strategies).

You May Also Like

What Does It Take to Democratize AI? Insights From the Field

Read More

Navigating Targeted Ad Bias With Responsible AI

Read More

Streaming With Dataiku Using Kafka & More

Read More

Enterprise Causal Inference: Beyond Churn Modeling

Read More