Validate

Before building there is a quiet phase that we almost always skip.

It is not recklessness. Listening looks too much like doing nothing, and when the project is yours, doing nothing is frightening. In the first days a strange urgency turns up, the sense that if you do not start now somebody will get there first. It is almost never true, and that hurry gets paid for later.

It does not help that this phase produces nothing you can show. No screens, no visible progress, nothing to tell anybody you did this week. Working on your own that weighs more than it sounds, because the only person measuring your progress is you.

What I am actually looking for

What I am after in this phase is not an opinion about my idea. Opinions about ideas are kind by default: people say it sounds good because saying otherwise would be rude. You take that sounds good home and file it as evidence, when it is not.

What I am after is what somebody is already doing without me. By hand, awkwardly, with nobody having asked them to. That awkwardness is the most reliable signal I have found, because nobody bothers solving something they do not care about.

The difference between the two is enormous. An opinion tells you what somebody believes they would do in a hypothetical situation. A workaround tells you what they already do, at what cost, in the real one. The first is forgotten the moment the conversation ends; the second has been going on for months.

What to ask, and what not to

The question that does not work is would you use this. It is hypothetical, it is flattering, and it comes back yes almost every time. Asking about the future is not much better: would you pay for this gets an honest answer that is still useless, because nobody really knows what they will do in three months.

The ones that have worked for me all point backwards, at what already happened. When was the last time you had this problem. What exactly did you do that day. How long did it take. What had you tried before. Why did you stop. What did it cost you, in money or in hours.

Those are questions about facts, and you cannot be polite with a fact. If somebody cannot remember the last time they had the problem, it does not happen as often as you thought, and you just found that out in thirty seconds.

Where to find those people

You do not need a representative sample or a panel. Eight or ten conversations and the pattern shows, and from there the same phrases start repeating.

The places where people already complain are the most useful: forum threads, reviews of similar products, trade groups, replies to posts on the subject. People write there without being asked, which is exactly the condition you are looking for. A spontaneous complaint is worth more than an arranged interview, because nobody writes one to be nice.

It is also worth talking to somebody who decided not to solve the problem. That is the most uncomfortable conversation and the most informative: they will explain why living with it was cheaper than fixing it, and that reasoning is exactly what you are competing against.

The problem with listening to yourself

There is a part that is hard to admit. Your own enthusiasm weighs more than anybody else opinion, and if you work alone there is nobody to offset it. This is where I trust my own judgement least, not because I cannot observe, but because I want the answer to be yes, and wanting something changes what you hear.

The only thing that has helped is writing down what I expect to hear before the first conversation. If what I then hear matches what I wrote a little too closely, I get suspicious of myself. And whatever turns up that I had not anticipated is usually the most valuable thing in the whole round.

The mistakes I have made here

The most expensive was talking only to people like me. It is the comfortable option: contacts from the trade, people who get the premise immediately and improve on it while you talk. The problem is that they hand your own thinking back to you and you read it as confirmation.

The second was explaining the idea before asking anything. The moment you describe what you plan to build, the conversation stops being about that person life and becomes about your proposal, and from there all you get is opinions. I now say what I am working on at the end, when it cannot contaminate anything.

The third is subtler: writing notes up neatly. Rewriting what somebody said in your own words organises it and strips out everything that did not fit your framing, which is usually the only part that would have changed your mind. I write the sentences down as they came out now, however untidy.

What I do with what I learn

At the end of a round I look for two things. The first is the workaround that repeats: the improvised system several people have built for themselves, almost always in a spreadsheet or a notebook. That tells me which part of the problem hurts enough for somebody to have already spent time on it.

The second is the sentence that repeats without me having suggested it. When three people describe the same situation in similar words, I have both the problem and the way to name it. That sentence usually ends up as the text on the home page, and not by accident: it is the only part written in the language of the person who has the problem.

How I know I am done

I know I am done when I can talk about specific people instead of profiles. Not freelance designers, but somebody with a name who last Tuesday did this in this way and lost an afternoon to it.

This phase is shorter than it looks: there comes a point where you are no longer learning anything new, you are looking for somebody to give you permission. When the third conversation in a row tells you nothing you did not know, it is over.

And if after ten conversations nobody has the problem in any recognisable form, that is a result too. It is not pleasant, but you found out in two weeks what would otherwise have taken six months of building.

Questions people ask about this

How many conversations do I need before building?

Eight to ten within one profile. What marks the end is not the number, it is running out of new things to hear: when the third conversation in a row adds nothing, you are done. If you are covering two very different profiles, count eight of each.

What if I do not know anybody with the problem?

Go where people already complain without being asked: forum threads, reviews of similar products, trade groups. A spontaneous complaint is worth more than an arranged interview, because nobody writes one in order to be nice.

Can I skip this if I have the problem myself?

Having it yourself is a good starting signal and a poor test. It tells you at least one case exists, not how many more there are or whether it hurts them enough. Three or four conversations will tell you whether you are a case or a pattern.

How do I tell a real signal from my own enthusiasm?

A real signal is something that already happened without you: a workaround built by hand, time invested, money spent. Enthusiasm feeds on sentences in the conditional. If what you have is intentions rather than facts, you do not have a signal yet.