A reminder to myself: one problem, said in one sentence, that people already feel and have already paid to fix. If I have to teach them the pain, it is not a startup problem.
A reminder to myself: one problem, said in one sentence, that people already feel and have already paid to fix. If I have to teach them the pain, it is not a startup problem.
This is a reminder to myself, written in September 2026. None of it is new. I still don't do it from the beginning, and I keep relearning it the slow way.
The problem has to be crystal clear, and it has to be one problem. Not a platform, not three related pains, not "it also helps with". One sentence that the customer would say in their own words.
I am still struggling with this. At my last company we built three products in under two years: a marketplace, a store builder and a catalogue. Each one made sense on its own. Three products is not one problem.
If the other person needs it explained before they feel why it hurts, that is the answer. The pain is not there. They may agree with me, politely, once I have explained it. Then nothing happens, because a pain you do not feel every day is not a pain you spend money on.
A big company can afford to educate a market. It has the budget and the years. A startup does not have that luxury. It has to go where people already know the pain, already talk about it, and are already annoyed enough to act.
On the latest ecommerce product I am testing, I added one line to the ideal customer. They have already tried to solve it. They bought a tool, hired someone, or hacked a fix together with Claude or Cursor. It still isn't solved.
That one line does a lot of work. It means they have already felt the loss, in money or hours. They already believe the problem is worth paying for, so I don't have to argue for it. And they have a comparison in their head: the tool that half worked, or the freelancer who cost too much.
After that, my job is smaller and clearer:
This filter makes the market smaller, on purpose. I am still testing it, so I have no results from it yet, and I am not claiming it works. It is the rule I wish I had started with. I will write the follow-up when there is something to measure.