A lot of startups are a solution in search of a problem. You build something, you fall in love with it, and you are certain everyone needs it. Then you go looking for the people with that problem and find out there aren't many. Or they have it and won't pay a cent to fix it.
I learned this the expensive way, with my own company.
When we built Caribu, we were sure we knew our customer. Traveling parents. The consultant who flies out Monday and wants to read a bedtime story to their kid from a hotel room. Aspirational, relatable, the story everyone nodded along to. People told us, to our faces, that this is what they wanted.
Then we looked at the data. The traveling parents weren't using it. We could have built every feature they asked for, made the whole thing free, and they still would not have shown up. A traveling parent still lives with their child and sees them pretty frequently. The pain was real, but it was small.
So we ran a proper product-market fit survey, the one Rahul Vohra of Superhuman wrote up. (If you do nothing else this week, go read that article.) We talked to the people actually paying us. And the customer turned out to be grandma. We called them glammas, glamorous grandmas. A grandparent barely gets to see the grandkid. A normal video call with a five-year-old falls apart in ninety seconds. For grandma, the gap our product filled was enormous, and she was having forty-five-minute calls on it.
Here is the part that matters. Grandma had almost no price ceiling. We could have charged her nearly anything, because we were solving a problem that genuinely hurt. We kept the price normal, but we realized we could have sold thousand-dollar lifetime subscriptions and she would have said yes without blinking.
That is product-market fit. Not "people say this is a nice idea." An expensive problem, and a specific person who will pay to make it go away.
The framing I give founders now is "vitamin" versus "antibiotic". A vitamin is nice. You take it to prevent things, and if you skip a day you are fine. An antibiotic is not optional. You are sick, you need it, you pay whatever it costs. For traveling parents, we were a vitamin. For grandma, we were an antibiotic. You want to be an antibiotic.
The trap underneath all of this is emotion. In education and much mission-driven work, we get emotionally attached to the problem and the people we want to help. I have been the founder reciting the five pillars of reading like scripture. But the question that actually keeps the lights on is colder than that. Who has this problem badly enough to pay, and why them?
If your answer is "this is important, and people need it," you do not have a business yet. If your answer is "this specific buyer is mandated to spend two hundred thousand dollars on this exact thing, and we do it cheaper and faster," now I am listening.
So before you build another feature, go find your grandma. Figure out who has the expensive problem, talk to the people already paying you, and follow the money instead of the story you wish were true.
Next, the flip side of all this: how to tell whether you actually have a business, or just investors keeping you alive.
A note on how this got written: I used AI to help me write this. The stories, the opinions, and the scars are all mine. I just have a very good robot assistant who helps me get them out of my head and onto the page. Which is more or less the whole point of what I do now.