MVPs Reveal Preference

Paul Samuelson introduced the concept of “revealed preference” in a 1938 article entitled A Note on the Pure Theory of Consumers’ Behavior. It means the preference implied by a consumer’s behavior in the market. It’s the economist’s broadening of putting your money where your mouth is.

What does this have to do with MVPs? An MVP is a tool to reveal preference.

I once worked with a team that wanted to build a mobile app to help people rate their phone carriers. The team started with a paper prototype MVP. They showed it to people and asked them what they would say about their carrier. People responded with great enthusiasm and told them all sorts of things positive and negative about the carriers.

But team learned only “stated preference”, the preferences expressed in response to questions or a survey. They didn’t learn revealed preference, how people behave when given the choice.

So I sent them out again. This time they introduced the prototype and simply handed it to people with a pencil and asked them to fill it out. The response was completely different. Almost everyone declined to participate. The revealed preference was to avoid the app, not to engage with it.

When building a product, how do you learn how users will interact with it? Will they in fact adopt it? How will they behave when they use it? Will they prefer it to the alternatives? How much is it worth to them? These are all questions of revealed preference.

A successful product engages users, changes their behavior, and improve their lives. Users may say one thing and do another. What matters is their behavior, their revealed preference. And the proper use of an MVP is to reveal that preference. If you genuinely understand users’ revealed preference, you can build a product that serves them well.

Jeff Patton has a good video on the history and interpretation of the term Miminum Viable Product, and says “Product” really means “Test”, an experiment to test some hypothesis. More recently Gagan Biyani took up the same refrain. Regardless of what you call it, focus your hypothesis on revealed rather than stated preference, and build an MVP to test it.

An MVP is a tool to reveal preference.