The first ten users — how to get them before there is a product

Founders ask us when to start thinking about users, and the answer is almost always "before you started building". Not because it is nice to do, but because ten people you have spoken to change the scope — and scope is what determines how much time and money the project costs.

The common mistake is treating this as marketing. At this stage it is not marketing. You are not trying to persuade anybody to sign up; you are trying to find out whether the problem you identified bothers anyone enough that they have done something about it.

Why ten rather than a thousand

Because you are not measuring, you are learning. A thousand sign-ups from a landing page tell you the headline worked; ten conversations tell you why those people are still solving the problem in a spreadsheet and what would make them stop. The second determines what to build. The first does not.

Usability research has shown for years that most interface problems surface with only five users, and the logic holds at the earlier stage too: after seven or eight conversations you start hearing the same things again, and that is the signal you have enough to decide.

Where the ten come from

Not from advertising. They come from four places, in descending order of efficiency, and all of them mean talking to named people rather than broadcasting at an audience.

1. People you already know who have the problem. Fastest and most biased — take what they tell you with your eyes open.
2. Places the audience already complains: professional groups, forums, replies under posts on the subject.
3. Introductions through introductions — asking each person you speak to for one more name. That doubles the list with no marketing effort.
4. Customers of an adjacent product: somebody already paying for something nearby has proved they spend money in the category.

The third is the one founders forget. One question at the end of every conversation — "who else do you know dealing with this?" — usually produces half the list, at no marketing cost.

How long it takes

Two to four weeks, and most of that time goes on scheduling rather than on the conversations themselves. A conversation runs twenty to thirty minutes, and two a day fits in a week without swallowing it. Founders who get stuck for two months have almost always built a landing page and waited for people to arrive, instead of approaching specific people by name.

Do not wait for all ten before drawing conclusions either. A pattern starts showing after three, and by five you can change the questions. The last conversations on the list are worth less precisely because they confirm what you already heard, and that is itself the signal you can stop.

What to ask

Not "would you use a product that does X". The answer to that is always yes and it is worth nothing. Ask about the past, because the past happened and the future is a guess.

• When did you last run into this, and what did you actually do about it?
• How long did it take, and what did it cost you?
• Did you try to solve it? What happened?
• If you didn't try — why not? That is the question that reveals whether the problem bites hard enough.

What to do with what you heard

Convert it into one question version one has to answer, and delete from scope everything not required to answer it. That is exactly the work described in how to define an MVP that doesn't blow up, and the conversations are what turn it from a guess into a decision.

And the most valuable part: those ten are also your first users. Somebody who gave you twenty minutes before there was a product is nearly always willing to try version one and tell you what does not work. That shortens the loop between launching and learning from months to days.

If you are already at the technical planning stage, it is worth reading how long an MVP takes too — your own response time to questions is the variable that decides more than any other.

One recurring mistake

The mistake we have watched most often is talking only to people who already agree. A founder approaches five friends from the industry, all five are enthusiastic, and they come away convinced. Those five were selected precisely because they resemble the founder, so what they report is the founder's own problem rather than the market's.

The fix is simple and unpleasant: make sure at least half the list is people you did not already know. They give less flattering answers, cancel more meetings, and tell you things you did not want to hear — which is exactly what makes the two weeks worth spending.