Every startup founder that’s not living under a rock should know, by now, that they’re supposed to “validate their market”, though probably many aren’t entirely clear what it means1.
This is true for “NFT startups” as much as for the more “traditional” kind. Whether you’re operating in web3 or dreaming of running your own taco stand, the principles are the same.
Wanna know the essence of market validation for early scrappy startups that just don’t have the time or inclination to do committees and reports an “research” and all those things? Read on.
Step 1: you’re wrong!
You can’t do something if you don’t know what it is. So what is market validation?
Well, the essence of it is pretty simple. It’s the answer to the question “does anyone want what I’m building?”
Or, more specifically, because this is market validation and not just audience validation: do enough people want to pay for what I’m building?
Now, for most first-time founders, the answer will seem obvious! Of course people want to pay for what I’m building. Look how useful it is! Once the market realises my genius, they’ll be falling over themselves to give me money!
Sadly, the world doesn’t work quite like that.
The first step to market validation is a bit like the first step of a 12-step programme: admit that you have a problem. The problem is: you have no idea if anyone wants to buy that thing you’re wanting to build. You have no idea who they might be. You don’t know whether they want to use it. You don’t know whether they want it enough to pay for it. You don’t know how much they might want to pay for it. And even if you pass all those hurdles, you don’t know how many of those someones there are.
If you think you know, you’re actually further behind than someone who doesn’t know. It might be frightening to admit how little we actually know, but when building a startup, it’s pretty essential to let go of those things we believe we know, and go find out the truth. As Mark Twain put it, what gets you in trouble isn’t what you don’t know, “it’s what you know for sure that just ain’t so”!
So, always begin from the perspective that you don’t have a god damn clue what the market actually wants. Then start looking for actual data, rather than beliefs driven by motivated reasoning2. Which brings us to step 2: how to get the data.
Step 2: Get that data!
Now that you have achieved the Socratic ideal3 in the field of market validation, time to actually get to know something.
The mistake many entrepreneurs make here is they do one of two things, both of which are wrong:
- They talk to friends and other random people, tell them their brilliant idea, and ask if it’s a good idea, and by the way, how much would you pay for it?
- Alternatively, they do some “market research” and find some big numbers that get them feeling butterflies in their stomach (“The total market is $40b a year! Imagine if we just get 1% of that!”).
Both of those approaches are bullshit and won’t help you.
Let’s break down what we want to learn:
- We want to know if anyone wants to use our product/service
- We want to know if they’re willing to pay anything for it
- We want to know how much they’re willing to pay
- We want to know how many of these people there are
The only way to get a definitive answer to all those is to build the business and run it for a few years. However, that’s a very expensive way to learn, especially because at this stage, we’re probably wrong about any assumptions we have on this. If we build a product now, it’ll almost certainly be the wrong product.
What is the minimum amount of work we can do to learn anything about these questions?
Well, first of all we can immediately strike out question 4, because if we don’t have a clue if anyone wants to pay for this, then we also don’t know who they are, and so we can’t figure out how many of them there are. This rules out the “market research” approach – those big numbers can’t help us. We need to start at step 1.
Ultimately, the only real proof that people want to use and pay for the product is to see them use and pay for the product. If it is possible to achieve that from where we are now, for example, if we have some kind of prototype out, then we can beeline for that. But if we don’t have anything yet, there are still some kinds of questions we can ask that will help us know a bit more about whether it’s worth building anything at all. But they’re not the questions founders (especially first time) typically ask.
The problem with questions like “would you use and pay for this product?” is that they give us bad data. Most people will care more about protecting our ego than about telling us the truth. So they’ll just lie to our faces. They don’t do that out of ill will, it’s just uncomfortable to tell someone that their idea sucks. But there is a solution to this, which is to learn to ask the right questions.
When we ask about hypotheticals, like “would you do this…” or “will you consider doing that…” we get bad data. But when we ask about actuals, people usually don’t lie about those.
So, here are examples of good questions we could ask:
- When is the last time you’ve encountered <problem that I’m solving>?
- How much trouble did that cause you?
- How much did it cost?
- Have you ever paid to try and solve this?
- How often do you encounter this problem?
- Do you know anyone else in your industry who has that problem?
You get the idea, I hope. These questions are all focusing on the actual, real, lived experience of the person we’re talking to. From these, we are likely to get truth out of them. Ask these types of questions of enough people, and we can start to piece together a sense of whether there is in fact a market for this solution we’re building.4
But there’s one more thing we need to remember when asking those questions: we have to behave like scientists, not like politicians.
Politicians are trying to convince people to get them to do something. (Good) scientists are trying to figure out the truth, and to be good scientists they consider all the possible hypotheses.
When asking these questions, I find it helpful to remind myself, before the conversation, that I’m almost certainly wrong about some critical aspect of my idea, that continuing to be wrong about it will possibly kill my startup, and this conversation is a golden opportunity to figure that thing out before it costs me months of wasted effort.
So, start each conversation with the assumption that you don’t know.
Step 3: Build an “MVP”, and do it again, and again, and again…
Everyone in the world of startups knows what an MVP is. Except they don’t. Most people misunderstand what “MVP” actually means, thinking it is an early prototype of the product they have in mind.
The term “MVP” itself is somewhat misleading. When people hear it spelled out, “Minimum Viable Product”, they mostly focus on the “Product” part rather than “Minimum Viable”. The interesting bit, which makes this a powerful concept, is actually those first two words.
“Minimum Viable” implies an intent. Minimum viable… for what?
Opinions about this may differ, but most people who have been in the startup world for a while agree that the core purpose of an early startup is learning. Your job as a founder is to get this fledgling organisation to learn and adapt so quickly that it ends up succeeding, against all odds, and survives. This also provides the direction for the viability:
The next MVP you should build is whatever enables the most learning.
In some cases, this may be a prototype. But in the vast majority of cases, it isn’t. In his seminal book The Lean Startup, Eric Ries gives the gut-wrenching example of having his team spend months building a chat integration module, only to find, upon deploying it, that not a single person clicked the button to download it. Not one.
In this case, the MVP was a button. That was what needed to be built to learn that no one will click the button.
Now, it’s not always that simple. Sometimes you do have to build something a bit more tangible to learn whether or not people want it. But even in those cases, your “MVP” design should be ruthlessly cut down to be the bare minimum you need in order to learn the next most important thing about your market.
The essence of the MVP is that it is something which enables you to observe some actual behaviour of your target market and learn from that behaviour, rather than continue speculating based on people’s words.
First of all, if you’ve enjoyed this article so far, please consider subscribing to my Substack to be sure to be notified of future articles.
Once we’ve built MVP after MVP after MVP, and learned many things about our market… when can we say we’ve “done” our customer validation?
Well, sorry for the cold shower, but the answer is: never.
Customer validation is not a job that we can complete. Remember:
The only way to get a definitive answer is to build the business and run it for a few years.Me, a few paragraphs earlier.
Actually, I’m wrong even in that statement. Customers are not a fixed target, especially these days. They change, their needs evolve, new products come onto the market, old products die out, behaviours adjust, old people retire and new people enter the workforce. Business life, like life, is a never-ending Darwinian battle where only the fittest survive.
Customer validation never finishes. If we want our business to last, there should always be some of the business’s attention dedicated to figuring out customer’s needs and testing whether they’ve changed.
That being said, there is a point where we can say the customer validation is “good enough” for specific things.
For example, one common milestone is “good enough to speak to investors”. What’s good enough? Well, that depends on the investors. Seed angels will want to see something. Series A VCs will want to see something else.
And there’s also “good enough to start building”. Well, there, again, it really depends on what we’re building, how expensive it is, etc. We will want more validation before committing to a two year project5 than before dedicating three days to building a feature.
Customer validation never ends. But at some point, it becomes an activity that we do alongside with running the business. How and when that happens really depends on the specifics, but often, the gears shift when a magical state called “product market fit” is achieved. At that point, customer validation necessarily slows down because we simply no longer have the time for it, because we’re too busy scaling the business.
Perhaps that will be a topic for another article. If you want to know when it comes out, you should subscribe here.
though they might wrongly think they do…↩
Motivated reasoning is when you really want a certain result to come out of your reasoning, and so you selectively ignore mental paths that might contradict your desired conclusion.↩
Socrates claimed he knew only one thing: that he knew nothing.↩
If you want to learn more about this type of questions, I strongly recommend Rob Fitzpatrick’s “The Mom Test”, still the ultimate, brief, easy to read, humorous manual on how to ask questions so good that even your mom will give you good data.↩
Does anyone still do this in 2023??↩