Context for recommendations – The 200 words project

When YouTube first introduced the “Recommendations” feature, it performed well but wasn’t impressive. Soon, the YouTube team added a simple tweak – below their “recommended for you” list, they added context as to why they recommended the video.

This completely changed the dynamic – where previously users saw recommendations that seemed to not make sense, they just said “YouTube recommendations suck!” and moved on. Now, when they had context, they understood why YouTube recommended a particular video. Thus, click throughs went up by ~20% and also created a great dynamic (for YouTube) in that, when the user found a bad recommendation, they blamed themselves and not YouTube.

So, if we’re aiming to create recommendations for our customers based on their past behavior, let’s consider providing context for better impact.

Netflix’s Product Managers have clearly followed suit. These are the shows recommended to me because I watched David Attenborough’s “Life” series – really helpful and adds so much to my experience.

Context for Recommendations

I’m generally a fan of experimenting with additional context when it provides a look inside your algorithmic black box. I think it makes your product seem more human to its users and feels like your algorithms are working on behalf of the person and their interests rather than just treating them as a row in a user database. – Hunter Walk

Source and thanks to: VC and ex-YouTube Product Manager Hunter Walk’s Blog