What question is your MVP asking?

N. Taylor Thompson wrote a blog post on building Minimum Viable Products that I was told was interesting but difficult to get through, so here's my attempt at an easier expression of the ideas:

----

A business model is a set of assumptions.

Minimal Viable Products (MVPs) are the experiments used to either validate or invalidate the assumptions in a business model.

Every MVP should be asking a question.  The answers affect the next step you take.

What question is your MVP asking? What does the answer tell you?

If the answer doesn't affect your subsequent behaviour, why did you bother asking the question? That is, if the next step doesn't change depending on whether the MVP succeeds or fails, then the MVP was a waste of time.

Different types of MVPs provide different types of answers.

Your MVP can be worse than the final product

That is, not as slick, handicapped in some way, but still contains the essence of the solution.

SUCCESS: Indicates that you have a strong problem-solution fit

FAILURE: Could be because the concept is flawed OR it could be because the product isn't good enough yet.

N. Taylor Thompson calls this a Validating MVP.

Your MVP can be better than the final product.

This is the essence of the Concierge MVP approach. Create a better product at an unsustainably high cost by personalising for every customer.  In essence, deliver magic and see if enough people are interested.

SUCCESS: Indicates that you have a market for your product, though you need to work out how to be profitable.

FAILURE: Time to pivot or move on.

N. Taylor Thompson calls this an Invalidating MVP.

Start with a Concierge MVP, follow-up with validating the details of the business model

N. Taylor Thompson suggests that for any situation with significant market risk, the most effective approach is to start by asking whether there is a market if the solution was magic, that is, with an MVP better than the final product (aka "invalidating MVP"). This is a more efficient way of detecting whether you're solving the right problem.  Only if that is successful should you worry about validating the other assumptions of the business model using MVPs that may be worse than the final product (aka "validating MVPs").