Your ICP is a job title and a company size. That’s a search filter, not a profile.

Ask a team who their ideal customer is and you get a job title. VP of Sales. Head of RevOps. Sometimes a company size bolted on. Fifty to two hundred employees, Series A to B.

That’s not an ideal customer profile. That’s a search filter.

An ICP is a description of who succeeds with you and why — specific enough that two people reading it would disqualify the same lead. Most ICPs fail that test. They’re broad enough to justify talking to anyone, which is the same as describing no one.

The profile and the person are two different things

The ICP is the company. The buyer is the person. Teams collapse them, and then wonder why the messaging lands with the economic buyer and dies with everyone else in the room.

You need both. The profile tells you which companies are worth your time. The buyer, the persona, tells you who to talk to inside them, what they care about, and what will make them move. A perfect-fit company with the wrong person in the room is a lost deal that looked qualified the whole way down.

Where the real profile comes from

Not a workshop. Not a whiteboard where the founders decide who they wish they sold to.

It comes from two places, and you need both. The first is research intelligence — firmographic and behavioural data about who actually bought, who renewed, who expanded, and who churned. The pattern is already in your CRM if the data is clean enough to read. Which companies closed fastest. Which stalled. Which came back. That’s the skeleton.

The second is conversations, and this is the half everyone skips because it’s slow. The data tells you what happened. Only a conversation tells you why. Why they bought when they did. What they nearly bought instead. What the meeting before the meeting sounded like. You cannot infer the trigger from a closed-won date. You have to ask.

Do both and the profile stops being a guess. We’ve classified a whole database this way — and the finding, more often than not, is that most of the list was never qualifiable in the first place.

Talk to the users, not just the signatures

Here is the part that separates a real ICP from a persona deck.

Everyone interviews the decision-maker. The person who signs. And they should — but the decision-maker is usually the person who knows the least about the daily reality of the problem you solve. They approved it. They don’t live in it.

The people who live in it are the users. The rep who has to work in the CRM you’re fixing. The ops person who runs the report. They can’t sign the contract — and they can absolutely kill it. A tool the users hate doesn’t get renewed, no matter who signed. And a tool the users love gets championed upward in a way no salesperson ever could.

So talk to them. Not to qualify them — they can’t buy. To understand them, because they’re the ones who make the decision-maker’s decision for them, quietly, before the decision-maker knows there’s a decision to make.

What motivates them isn’t what motivates the buyer

The economic buyer is motivated by the number — the risk, the board slide, the cost of doing nothing. Fine. Sell that to them.

The user is motivated by their Tuesday. Whether this makes their actual job less annoying. Whether they’ll have to think about how to use it, or whether it’ll just work. Sell them the Tuesday. Two different motivations, two different messages, same account — and if you only have one message, you’re talking to half the room.

How you know it’s working

A real ICP does something a search filter never does: it lets you say no. It disqualifies good-looking leads early, on purpose, so your team spends its time where it converts. If your ICP has never made you walk away from a deal that fit on paper, it isn’t doing its job. It’s decoration.

The test isn’t whether you can describe your best customer. Everyone can, after the fact. The test is whether the description tells your team who to ignore on Monday morning — before the deal, not after the loss.