Every Company Needs an ICP. Even Yours.

If I asked three people at your company to describe your best customer, would I get three different answers?

---

You need an Ideal Customer Profile.

“Not us,” the founder says. “We don’t turn away revenue. We sell to whoever’s buying. Picking one type of customer sounds like a good way to leave money on the table.”

But an ICP was never about who you refuse to sell to. Nobody’s asking you to turn away a check. It’s about where you point your limited targeting and resourcing effort on purpose, instead of wherever the next inbound happens to come from.

You can still take the revenue from outside your ICP – and many companies with sharp ICP definitions do just that.

But you shouldn’t be building your marketing, your sales motion, or your roadmap around ‘whatever comes in.’

Most companies don’t actually have an ICP. They have a persona.

Ask most founders for their ICP and you’ll get something like “tech companies, $5 million to $75 million in revenue.” That’s directionally true, but it’s also not specific enough to do anything with.

That’s a persona. It’s the kind of filter you’d punch into a LinkedIn ad campaign: industry, headcount, revenue band, maybe a job title. It describes a demographic. It doesn’t describe a company.

A real ICP understands how the business actually runs. What tools and systems they’re already using. What their internal process looks like for evaluating and buying something like what you sell. What’s actually causing them pain right now, in their own words, not yours. How they’ll consume and implement the solution once they’ve bought it.

“Industrial tech companies, $5 million to $75 million” is a filter. “Industrial SaaS companies built on Red Hat, still running deployment through a manual release process” is an ICP.

The second one tells you exactly who to look for, what they’re already dealing with, and what to say to them that a demographic filter never could.

The more specific you get, the better it works.

This feels backwards to most people. Narrower targeting means fewer opportunities, not more of them, right?

Wrong. A vague ICP gives your team nothing to aim at. Marketing writes copy that could apply to half the market, so it lands with none of it. Sales prospects broadly and qualifies late, burning cycles and time on companies that were never going to be a fit. Nobody can tell you, with any confidence, how a deal is going to close.

A sharp, specific ICP does the opposite. It tells you exactly which accounts to go after, informs what language will actually land, and predicts what to expect once you’re in the room. Specificity is what turns targeting from a guess into a system. Aim small, miss small.

Your ICP should evolve as your company does.

The ICP that’s right for you today probably isn’t the one that gets you to the next stage. That’s expected. It’s supposed to move.

One way to track it: whether you’re still proving the problem is real (problem-market fit), proving your product actually solves it for that problem (product-market fit), or proving you can deliver it repeatedly at scale (platform-market fit). Your ICP at each stage should look different, because what you’re actually trying to prove is different.

The plainer version: startup, emerging, established, enterprise. A startup’s ICP is usually whoever gives you the clearest signal fastest. An enterprise’s ICP is usually the account worth the much heavier cost of landing and keeping. Neither one is wrong. They’re answering different questions at different moments.

Can’t get that specific yet? Start with the problem instead of the company.

Some of this is harder to do early. If you’re a startup, a service business, or a consultant, you may not have served enough of the right customers yet to describe a sharp, specific company profile with any confidence. That’s normal. It doesn’t mean you get to skip the exercise.

What you can usually describe, even on day one, is the problem. Call it the Ideal Client Problem instead of the ideal client profile. Not the industry, size, or tech stack of the company you serve best, but the exact situation you solve better than anyone else. What’s the inflection point that you can guide them through?

A consultant might work across a dozen industries and every company size in between, so a firmographic ICP won’t hold together. But the actual problem, the specific situation that keeps showing up right before someone hires you, usually is consistent. “A founder still personally approves every piece of content six months after hiring a marketing lead” is a sharp, specific problem, even if the company behind it sells industrial software one month and specialty food the next.

Anchor on that instead. It gives you the same benefit a specific ICP gives a more established company: something narrow enough to build messaging and targeting around, instead of something so broad it fits everyone and lands with no one.

As you serve more of the right clients, the Ideal Client Problem usually starts revealing the sharper company-level ICP underneath it. The problem was never a replacement for the profile. It’s just where you start when you don’t have enough history yet to see the profile clearly.

How many ICPs should you have?

As many as you can support. That number is smaller than most think.

Every ICP you support needs its own messaging, its own proof points, its own sales talk track, motions, and usually its own version of how you onboard and support them once they’ve bought. That’s not a one-time cost. It’s ongoing, for every ICP, for as long as you’re targeting it.

I watched this play out firsthand. Years ago I was inside an agency right after it had gone through an acquisition. Two client rosters got merged into one, and on paper it looked like a win. More logos, more industries, more revenue on the table.

It wasn’t. A meaningful chunk of that combined roster was unprofitable the moment you accounted for what it actually cost to properly support them. We were fast for some clients and slow for others, strong in a few verticals and stretched thin everywhere else, because we were trying to be the right fit for more types of customers than we could actually serve well.

We picked the ones we could genuinely support and let the rest go, on purpose. Revenue grew rapidly in 6 months and the Net Revenue Retention rose almost 8 points. None of that came from adding more customers or services. It came from deciding we wouldn’t support all of them, and cutting accordingly.

Most companies can genuinely do this well for one to three ICPs. Past that, something gives, usually the thing you don’t notice slipping until the numbers show it.

Would your team’s answer to “who is this for” hold up if I asked three different people, in three different departments, right now?

Many companies already have the data to inform the work. They just haven’t pulled it apart by customer – or know what to look for.