Recommendation is about trust. Trust the data, not the math.
|Jean de La Rochebrochard||Aug 27, 2013|
Recommendation has been there for quite a long time on e-commerce, ad-targeting, applications for TV, Books, Restaurants, Video Games… And it has been often based on the same pattern: Statistic.
Applications using recommendation usually collect and analyse data to extract the best possible outputs toward its users. It’s based on calculation. While it instinctively seems to be the right way to do it, it’s actually not. The reasons are simple: it does not work without initial behaviour data and you can’t take friction into account as it would considerably and infinitely increase the complexity of calculation. you therefore tend to end up with a pile of bias which distorts the results and lead to bad/unrelevant/uncomplete recommendation.
And while we could argue for some time on whether or not statistic recommendation can be reliable at some point (yes, it is on Amazon for instance), we would still miss the point:
Recommendation is about TRUST. Your quest is toward your USERS’ SATISFACTION, and only. Often, your users use recommendation because they don’t know what they want. They need GUIDANCE, INSPIRATION. Statistic does not provide guidance until it knows them! And once it knows them, they find themselves stuck on a path built through their initial choices. The best example is Netflix. The comments below the recent interview from Wired “The Science Behind the Netflix Algorithms That Decide What You’ll Watch Next” reveal an obvious lack of trust and satisfaction regarding their discovery and recommendation service while they invest millions into it.