What users like vs what they might like : The recommendation algorithm dilemma.

We all kind of agree that Henry Ford did say, “If I had asked people what they wanted, they would have said fasterĀ horses.” ( I still haven’t found a legitimate source to verify the veracity of this quote. Let me know if you do.)

Doesn’t mean we shouldn’t listen to the our customers. (Consumers not merchants. SAAS models for B2B applications are out of the scope of this discussion.) We can understand that there’s a need for faster mode of transportation. Now, a faster horse could either be horses on steroids or an engine, that’s up to the guy who implements the solution. In this case Henry Ford was smart. We, obviously; don’t have doubt about that.

From this argument if I deduce that user does have general understanding of what they might want but not entirely sure about the final form of the solution, the app / platform should present different solutions to the users. It just might be your lucky day and they might like what they see; at which point, you should definitely have a feature to capture their satisfaction.

It doesn’t have to be an isolated (and usually boring) thumbs up button, which I find so lame and waste of time; but it could be a multifunctional “like” button which works both as a personal consent and a social validation.

So allowing users to explore through what they might like by broadening the algorithm to consider their “like profile” and suggesting what similar users might have liked 1000s of KMs away would allow the users to come across content they might have never thought of watching and actually enjoy it.

And probably this is one of major differences behind TikTok, PinDuoDuo’s social success vs a monotonous You Tube’s lonely feed. And as I always say the biggest part of the problem is also tied to the purpose : is your algorithm satisfying the users or making them click more on the ads.

If you take the purpose seriously, I think you already have the answer to the question . Remember, our duty as product managers, solutions architects, engineers and developers is to provide solutions.

I would love to hear the kind of solutions you have implemented lately.