An MVP Reveals Preference

Why does the term “Minimum Viable Product” or MVP confuse so many people? There are two common and distinct interpretations of the MVP concept. Each emphasizes a different part of the term, and neither focuses squarely on the most important point of MVP use, namely validated learning. Fortunately the field of economics has a well developed terminology called “revealed preference” that provides grounding for the MVP concept.

What are the common interpretations of MVP, and how does revealed preference help clarify them?

One interpretation of MVP focuses on “viable product” and emphasizes the need for the MVP to reach some desired threshold of feature completeness or refinement before exposing it to customers. People who think of the MVP as a prototype product gravitate to this interpretation, as do those who hold themselves to strict notions of brand alignment. This idea of MVP is less about discovery than about implementation.

The other interpretation of MVP focuses on “minimum” and emphasizes the need to push something out the door as quickly as possible. Those working to add small enhancements to existing products frequently use this interpretation. This interpretation is more discovery oriented and more experimental, and therefore closer to validated learning. Yet it can remain shallow. Do the customers click more if the button is in this or that location? What color should this button be? These sorts of tests teach you very little about your customers.

Some have suggested a new term such as “Minimum Viable Experiment” for the second interpretation to avoid the confusion. It’s unclear to me whether that will lead to more or less confusion. Let’s focus instead on validated learning as grounded in the language of economics.

Economists talk about two types of preference: stated preference and revealed preference. Stated preference is what you say in response to a questionnaire, or a survey, or in an interview. Revealed preference is inferred from your actions. Revealed preference is stronger than stated preference, because it derives from behavior and not merely from talk. It’s the economics equivalent of putting your money where your mouth is.

I may say that my preferred toothpaste is Crest. But when I go to the store and see a less expensive house brand I buy it. If I do that consistently, then my stated and revealed preferences differ.

One of my clients wanted to develop a mobile application for rating wireless carriers. The team developed a paper prototype and spoke to customers, showing them the prototype and asking “Would you use this to rate your wireless carrier?” and “What would you say about your carrier?” They received all sorts of interesting answers, and a high degree of interest in the product.

Readers experienced in customer interviews probably saw the danger upon reading “Would you…”. Of course they would! That type of question almost always yields a positive response. Stated preference.

After we debriefed on this experiment, the team and I talked about stated and revealed preference. They went out again to speak with customers. This time they handed the paper prototype to the customer and said “Please rate your wireless carrier.” The responses were very different. Many more people declined to interact with the prototype directly, as opposed to talking about it. This time the customers had to act. They had to put pen to paper and spend time recording their thoughts. Revealed preference.

This doesn’t mean that interviews or conversations are useless in discovery. Stated preferences may provide you with insights, but they are noisier and weaker than revealed preferences.

What does this imply about good use of MVPs?

First, it’s important that you ask for payment of some variety, so that the customer reveals a willingness to pay. The currency can be time, or information like an email address, or dollars. But without payment, it’s just talk.

Second, whenever possible have your customer interact with your prototype or product. Record their behavior and make your inferences from them. Don’t take their word for it.

Third, think of an MVP as an experiment to reveal customer preference. Stated preference is talk, and talk is cheap. Once you understand revealed preference and begin to treat your MVPs as preference-revealing experiments, you’ll get much more out of your discovery efforts.

comments powered by Disqus