How to run cheap and fast experiments before building a real MVP

The most common mistake of new teams is to dive straight into coding. It seems the sooner the product exists, the sooner you'll know if the market needs it. In reality, building an MVP is already an expensive and lengthy undertaking, and before it you can almost always test your key assumptions far more cheaply.

The first step is not to build but to talk. The method known as “The Mom Test” solves a common CustDev problem: out of politeness, people tell you what you want to hear. So you shouldn't ask “Would you buy a product like this?” — you need to ask about the past: “What did you do the last time you ran into this problem?” If someone has already cobbled together a clumsy homemade solution, that's a strong signal that the problem is real and painful enough.

Next comes the “fake door” test. It's a landing page that looks as if the product already exists: a description, a value proposition, a “Buy” or “Submit a request” button. Clicking the button leads not to a payment page but to a “Thanks, we're launching soon” page or a sign-up form. Such a test costs nothing in terms of development, yet it gives you real metrics: how many people from your target audience clicked and how many left their email. Y Combinator recommends exactly this approach — testing not people's willingness to buy the product “in principle” but a specific action: whether they're willing to hand over their contact details or money for something that doesn't exist yet.

For B2B products, there's a weightier option — a letter of intent, or LOI: a potential customer confirms in writing that they will buy the product on certain terms if it's delivered by an agreed date. This is no longer a guess but a signed commitment that you can show investors as proof of demand.

The main principle that all the sources repeat: at this stage, you test not the whole product but the riskiest hypothesis — the assumption that, if wrong, would kill the entire business model. If it isn't confirmed, you've saved months of development. If it is, only then does it make sense to move on to a real MVP.