How do I find my ideal customer profile?

Look at the people who already turned up rather than describing the people you wish would. Take your three best users, and for each one write down what changed in their world the week before they arrived - a new job, a headcount change, a tool that broke, a deadline, a price rise. If two of the three share that moment, you have found something you can act on: you can go and stand where people are at that moment. A demographic like 'B2B SaaS, 50 to 200 people' cannot be acted on, because it tells you who they are and never when to reach them.

What is the difference between an ICP and a buyer persona?

A persona is a description of a person: their role, their goals, sometimes a stock photograph and a name. An ideal customer profile is a decision about who you will and will not serve, and the useful half of it is the second one. A profile that excludes nobody is not a profile. If your ICP does not let you say no to a category of customer this week, it is a persona with a different job title.

Can I define an ICP before I have any customers?

You can write a hypothesis, and you should - but calling it a profile is how founders spend six months building for people who were never there. Before you have customers, the honest version names who you think it is, what moment you think they hit, and what would prove you wrong. Then you go and have ten conversations with people who match it. The value is in the disqualifying answer, not the confirming one.

Free tool · 4 minutes · no signup

Stop guessing
who it's for.

Most ICP templates ask you to describe your ideal customer. You already did that, and it's why the answer came out as "small businesses". This one starts from the people who actually turned up.

First, the only question that changes the method Have at least three people used it? Not signed up - used it, more than once, on purpose.

A demographic is not an ICP

"B2B SaaS, 50 to 200 people, Series A, VP of Ops" is a filter for a prospecting tool. It is not a profile, because it tells you who someone is and never when to reach them - and almost everyone who matches it is not in the market this quarter.

The people who bought from you did not buy because of their headcount. They bought because something changed in the week or two before they went looking: a person left, a site opened, a tool broke, an invoice arrived that was bigger than last year's, a customer complained twice about the same thing. That change is the trigger, and it is the only part of an ICP you can actually go and stand next to.

Why it's worth the discomfort of only three examples

Three real ones beat three hundred imagined ones. A survey of your target market tells you what people say about a hypothetical. Three customers tell you what somebody did with their own money, which is the only evidence that has ever been worth anything.
The trigger is findable, the demographic is not. If two of your three arrived the month they opened a second location, you now know where to be: the forums, suppliers, accountants and search queries that surround opening a second location. "Retail, 50 to 200 staff" gives you a list to interrupt at random.
The exclusion is the product decision. Knowing who it isn't for is what lets you say no to a feature request, a customer, and a channel - three decisions you make every week and currently make on instinct.

Then feed it forward

The answer this gives you is, almost exactly, question two of the positioning generator - the field that asks who feels the pain most sharply and warns you that "small businesses" is a census category rather than a segment. That tool assumes you already know. This one is where you find out, and the order matters: positioning written against a segment you guessed is just adjective soup with a narrower audience.

The honest limit of this tool

It reads what you typed. It cannot tell you whether your three customers are representative, whether you picked them because they're the ones you like talking to, or whether the pattern you're about to see is real rather than the shape of your own memory. Three is enough to form a hypothesis and nowhere near enough to be sure. The way to know is to find the next five customers by going to the moment this points at - if that works twice, it's a profile; if it doesn't, it was a story.

And if it tells you your three have nothing in common: that is a finding, not a failure. It usually means you have a product several different people can use for several different reasons, which is a real and survivable position - it just means you cannot yet buy attention efficiently, and you should stop trying until one of the reasons wins.

What to do with it this week

A demographic is not an ICP

Why it's worth the discomfort of only three examples

Then feed it forward