MVP? Is it a product or an experiment?

Whenever I write, talk or give a class about Product Discovery, I constantly emphasize on one of the main tools of the process of validation and continuous experimentation:

  • The MVP Test (“Minimum Viable Product Test”)

This tool has a fundamental problem with its name: basically the inclusion of the word “product”. But before we see why that is a problem, let us review what an MVP Test is and where this confusion comes from.

What is the MVP Test?

In essence, it is a prototype of the product, real enough to put it in front of the user and get feedback from it.

We want this feedback to be based on user actions and not opinions (what they do rather than what they say), so whenever possible we will ask them to take actions on the prototype and tell us their experience with what they are seeing so we can learn and make decisions for the final product.

Important note: by prototypes I’m not referring to the more traditional meaning we use in UX practices. While a Wireframe or Hi-Fi prototype can be used as an MVP Test, I also include in the definition other types of prototypes, ranging from landing pages built to test the value proposition to actual code that simulates a new feature and is shipped to a test device to show it to users.

Lean Startup – The source of this confusion?

While the term MVP has been around for quite a while, it became much more popular with Eric Rise’s book Lean Startup and the whole movement that was generated after its release.

In the book, Eric Rise introduces MVP as:

This can definitely cause confusion.

In fact, in a good post about MVPs, the same author starts with “Its power is matched only by the amount of confusion that it causes”.

An MVP must add value to the user.

A definition that I prefer is the one by Ash Maurya, author of Running Lean (a great book if you want to take the concepts of Lean Startup to a more practical terrain):

“A Minimum Viable Product is the smallest thing you can build that delivers customer value”

Instead, an MVP test (with special emphasis on the word test) is just something to experiment and learn. A landing page that communicates a value proposition to measure interest through a CTR is definitely not a product. It does not add any value to the user, it is just for us, the team trying to learn.

Minimum Viable Experiment (EMV or MVE)

In a spark of creativity (?) I decided to start calling these tests “minimum viable experiments”. 2 minutes later I run a Google search and saw that my idea was not that original:

The point is:

  • The goal of the MVP Test is to learn as quickly and cheaply as possible – “the minimum unit from which we can learn”
  • So most of the time (I would say 99%) what we build does not bring “marketable” value to the user
    • We build artifacts like prototypes, landing pages, fake doors, etc.
  • We do it in the context of an experiment to maximize our learning

So definitely what we are doing is not a product.

And since what we are doing is a minimal artifact that allows us to experiment, the term Minimum Viable Experiment sounds like a great description.

Do we dispose of the MVP?

No way!

Once we learn enough through experiments we want to build something “minimal” but that adds value to the user and allows us to continue learning to iteratively reach a complete product.

One of the great controversies is how “minimum” does it have to be in order to be “viable”. For me the answer is again in the definition of product: it has to add value to the user, and so I vote in favor of the definition of Ash Maurya mentioned before.

In short, I believe that the problem with “viable” occurs when people abuse the term to launch mediocre products.

Is there no more learning with MVPs?

Yes, there is!

Even with finished products you never stop learning. But in particular the experiments have limits to what we can learn, and without any doubt having a minimum product in the market will be a great source of new lessons to keep improving it.


In short, to get out of the discussion about MVPs, I will now call the MVP Test directly MVE: Minimum Viable Experiment – the minimum unit that allows me to learn without necessarily adding value to customers.

And I will limit the use of MVPs when the deliverable involves something valuable, as proposed by Ash Maurya.

I would love to hear your thoughts in the comments section!

Leave a Reply

Your email address will not be published. Required fields are marked *