From Customers to Pattern: Rethinking ICP
Why the customer you lost may tell you more about your ICP than the customer you won.
You raised the round. You have customers. Maybe even a few logos you're proud to put on the website. So you probably have a pretty good idea of who your customer is.
But do you?
Capital is a bet on your potential, not proof of your ICP.
And your first customers are proof that someone was willing to buy — not necessarily proof that you've found a repeatable customer pattern. The uncomfortable part is that the customer you lost might actually tell you more about your ICP than the one you won.
Your first five customers might not be your ICP
If you're a founder with 3 to 10 customers and you've decided “our ICP is B2B SaaS companies with 50–500 employees in the US,” you may be making a very expensive assumption. Because those five customers may have arrived for five completely different reasons.
All five paid.That is one thing they have in common — not a pattern.
Early customers are signals. They are not automatically your ICP. And this is where a lot of early-stage GTM advice gets the sequence wrong.
Before the ICP comes your PMF journey
At pre-PMF, you're not supposed to have everything figured out. You're trying to learn.
The questions you are actually answering
The job at this stage is not to scale what you think works.
It's to figure out what actually works repeatedly. That means founder-led selling, customer conversations, manual onboarding and changing your pitch are not necessarily signs that your GTM is broken. They may simply be part of the learning process.
The job changes as you move toward PMF.
Before PMF, your ICP is a hypothesis. At PMF, you start recognising patterns.
Learning:
That shift matters because ICP is a pattern, not a wish list.
So what actually is an ICP?
An Ideal Customer Profile (ICP) is a description of the type of customer for whom your product is most relevant, most valuable, and most likely to buy, stay, and grow.
An Ideal Customer Profile is not:
That last one is a filter. An ICP should help you answer a much harder question:
Where does our GTM have the highest probability of working?
That means looking beyond who can buy and understanding the circumstances in which they do.
Learning:
The pattern emerges from the intersection of these things.
The one who stayed isn't automatically your ICP either
This is where it gets uncomfortable. Your biggest customer may not be your best customer.
Which one would you rather have ten more of?
That's the question your ICP should help you answer. Not who gives us the biggest deal? — but:
The better question
Which customer pattern creates the strongest combination of value, willingness to buy, retention and economics for us?
Your ICP is not a customer you want to win
It's a pattern that emerges in the customer, and in everything around them.
You don't need to have that pattern perfectly defined on day one. Especially if you're pre-PMF. But you do need to know what you're trying to learn.
Because the goal isn't to create an impressive ICP slide. The goal is to find a customer pattern you can repeatedly win, create value for, and build a business around.
The pattern emerges when you stop looking at your customers as a list of company names and start looking across the conditions under which they become good customers.
What context are they in when they need you?
What problem are they trying to solve?
What trigger makes that problem urgent enough to act on?
How do they buy?
What value do they actually get from your product?
And just as importantly, what makes some customers easier to win, retain and grow than others?
The patterns across these answers are what start to define your ICP. And that's when ICP stops being a hypothesis and starts becoming a GTM advantage.
Comments