Revealed preferences and the data trap

Sometimes, I think the concept of revealed preference — the idea that someone may say they prefer something but when you actually observe them their preferences show something else — and it’s usage to make product decisions might actually be missing the forest for the trees.

And what I mean by that is a lot of tech product decisions are made based on revealed preference aka “This is what people are actually doing or want.” But sometimes that isn't good.

I'll give you an example. Someone may say “hey I actually want to learn about the environment or what's going on in my community.” But then when you actually observe their behaviors on a social media platform they may just watch hamsters in a wheel going round and around for hours. Now there's nothing wrong with that per se.

Watching a hamster and in a wheel go round and round is a revealed preference. Where the issue comes in is though that person may spend a lot of their attention on that particular thing we should never equate that behavior with what that person actually values or aspires to value. Which are often higher level, important things like community.

This takes us to the complaint I often see on social media, particularly Meta properties which is that people would rather see their friends and family’s content instead of suggested creator content. The interesting thing is that the revealed preference is for suggested creator content because that's what many or most users actually watch.

And this is where the reliance on revealed preference in making product decisions can be somewhat fraught. Because the original data said that the revealed preference is for a certain type of content, as a data-informed product manager or leader you start to over index on showing that suggested content because truthfully and factually users find it more engaging. Over time the scale tips toward suggested content even when friends and family content would perform equally well from a time spent on app perspective.

This ultimately may not be the best thing for the user experience because users are also being truthful in their complaints. And they're being truthful because they know internally and intrinsically that that content is probably not the best for them. And over time, the decision of what content to engage with is weighted heavily towards suggested content. This is where those complaints stem from. And as a business you should listen to that. There's an important reason why.

That user frustration is an opening for another company to create a product and experience that caters to that stated preference in a more engaging way. And if that's possible, that creates a real challenge to your business long term.