Why Your ICP Doesn't Work

Truly operational ICP definition makes you uncomfortable and risky. And that's what you should aim for.

It's common knowledge that, for companies to succeed, you need to have a clear Ideal Customer Profile (ICP). That's why most companies have it documented somewhere. They have a slide in their deck. They have a persona. They can tell you the industry, company size, and job title of their ideal customer.

And yet, the ICP rarely does any real work. It exists, but it doesn't actually constrain decisions.

This is why ICPs fail in practice. Because many companies create what I call the comfortable version of an ICP.

So what is this comfortable version?

Most ICPs are written to be reasonable, not useful. They describe:

• Who could buy the product
• Who has bought it before
• Who the company wants to see as customers
• Who the team doesn't want to upset by excluding

They're designed to include people, not exclude them.

An ICP that makes everyone in the room nod in agreement is usually a weak one.

"Yep, we captured it all!" everyone says.

And that's the problem at the pre-scale phase.

Because if your ICP doesn't force you to say "no" to some revenue opportunities, it's not really an ICP. It's just a market description - a description of everyone you hope to capture.

Now, when do ICPs start to break?

ICPs tend to break at a very specific moment: when the product starts working.

• Trial / freemium signups increase.
• Conversion improves.
• A few strong customers emerge and start telling others about you.

And suddenly, the pressure shifts from finding the right customers to not missing any potential customers.

That's when the ICP quietly expands:

"Well, we can also serve this segment quite well..."

"This use case actually converts pretty well too..."

"Let's not turn this channel off just yet..."

Each expansion feels rational in the moment. Each one probably makes business sense on its own. The ICP remains on paper. But quickly, it stops influencing reality and diluting your focus.

ICPs fail because they aren't operational.

The real problem with most ICPs isn't that they're wrong. It's that they aren't used as operational constraints. A functional ICP should shape actual decisions:

• Which features get emphasized in your messaging
• Which messages are used and where
• Which pages get built on your website
• Which traffic you actively want to attract
• Which support tickets you're willing to tolerate
• Which deals you intentionally ignore

If your ICP doesn't change behavior - if it doesn't affect how people make decisions day to day - it's just decoration.

In practice, many teams say they have an ICP while continuing to optimize for anyone who converts. Because inclusion feels safe, but exclusion feels risky.

Exclusion is genuinely uncomfortable. It means:

• Saying no to use cases that "kind of work"
• Ignoring geographies where you could get signups
• Not prioritizing customers who pay, but don't stay

For bootstrapped companies, this feels especially dangerous, because, let's be honest, every euro you get counts. Every signup feels earned. Walking away from demand feels almost irresponsible.

But the paradox that I've seen play out is this: the more constrained your ICP becomes, the safer growth actually gets.

Let me share what changed when we took ICP seriously. At the company I work with, the shift didn't come from rewriting personas or creating better slides. It came from making the ICP expensive. Expensive in the sense that:

• It cost us potential signups that we would have gotten otherwise
• It cost us to stop prioritizing an entire region
• It reduced our surface-level opportunity
• It forced genuinely uncomfortable trade-offs

We narrowed our focus:

• Narrowed our industry focus
• Narrowed our geographic focus
• Narrowed the job that the product was allowed to be good at

And something important happened. We didn't just attract fewer users overall. We attracted more similar ones.

That similarity completely changed the game. It helped us shift our focus in both product and marketing:

• We could make more focused improvements in onboarding because we knew exactly who was coming through.
• We could craft more focused messaging because we weren't trying to speak to everyone.
• We had better internal alignment because everyone knew who we were building for.

And we saw stronger conversion and better prospects for retention. We started acquiring more customers on annual subscriptions instead of monthly, too.

Those are the results you get when your ICP stops being descriptive and becomes directive - when it actually directs your decisions.

Now, here's another thing I learned to bring about this shift: the "No List" is actually as important as the operational ICP definition itself.

One of the most useful things I did was create a clear "No List." It's about having explicit statements like:

• We are not optimizing for XYZ usage
• We are not serving enterprise procurement needs right now
• We are not prioritizing this geography for the next 12 months
• We are not focusing on this particular use case

That No List did more work than any persona document ever had. Because it gave me permission to stop reacting to certain activities, requests, or urges. It gave us clarity about what we could ignore.

The reality is that ICPs fail when they don't hurt. If defining your ICP doesn't feel uncomfortable, it's probably too broad.

A strong ICP creates:

• Internal tension, at least until the results demonstrate the value of being focused
• Short-term trade-offs that feel risky
• Awkward conversations about "lost" revenue

But it also creates:

• Cleaner data that you can actually learn from
• Better decisions because the options are constrained
• Marketing that generates better-fit customers and starts compounding
• More satisfied and engaged customers down the road

ICPs don't fail because teams don't understand their customers. They fail because teams aren't willing to compress around the customers that matter most right now.

If you're in the pre-scale phase, stop asking yourself "Who could use our product?" Instead, ask:

• Who would we be willing to build this product exclusively for over the next 12 months?
• Which customers would we stop optimizing for if we had to choose?
• What revenue would we intentionally ignore to get clarity?

If your ICP can't answer those questions, it will fail in practice. And if it fails in practice, no amount of traffic will save you.