3 Ways to Make Better Product Recommendations, Like Netflix

Jul, 28 2014

You may have wondered how Netflix always seems to know what you like to watch. They know you so well that sometimes it seems like black magic. Suggested movies and shows appear at the bottom of your screen with movie titles you just can't avoid watching. So how does Netflix do this?

“75% of what people watch [on Netflix] is from some sort of recommendation” Netflix’s Research Director Xavier Amatriain said on Netflix’s tech blog.

Netflix is able to serve these relevant recommendations thanks to powerful customer clustering models. These predictive models analyze your viewing habits & match you with other customers who have similar tastes. By drawing from these clusters, Netflix is able to offer highly relevant recommendations. Companies like Shazam, Amazon, eBay, and StumbleUpon use similar models to make their own product recommendations.

The big question is how can marketers use recommendation engines to deliver better campaigns? We've put together a quick list of predictive models that you can use to make accurate product recommendations just like Netflix.

Let's get started.

1. Behavioral clustering

Behavioral clustering informs you how people behave while purchasing: do they use the web site or the call center? Are they discount addicts? How frequently do they buy? How much do they spend? How much time will go buy before they purchase again? This algorithm helps set the right tone when contacting the customer. For instance, customers that buy frequently but with low sized orders might react well to offers like ‘Earn double rewards points when you spend $100 or more.

2.Product based clustering

Product based clustering algorithms discover what different groups of products people buy from. The example above shows a category (or product) based segment or cluster. You can see people in one customer segment ONLY buy sweaters, whereas those in another customer segment buy different types of active wear products, such as outerwear, sportswear, swimwear and watches – but never kids’ clothes, intimates or jewelry. This is useful information when deciding which product offers or email content to send to each of customer segment.

3. Brand based clustering

Brand based clusters tell you what brands people like. Now you know what specific brands to pitch to certain customers. When a brand releases new products – you know who is likely to be interested. As you can tell from the example above, the algorithm has discovered that customers who like Tahari also tend to like Calvin Klein and Nine West, but would not be interested in Desigual or 6126.