How to find your ICP when you have zero customers
Standard ICP advice breaks down pre-traction. Here’s how to build a hypothesis fast.
The standard ICP advice is: look at your best customers and find patterns.
Great advice. Useless if you don't have any.
Pre-revenue, pre-traction, maybe pre-launch — you're building something for someone, but that someone hasn't shown up yet. You can't interview customers you don't have. You can't spot patterns in a dataset that doesn't exist.
So what do you actually do?
If you want the bigger picture of what ICP is (and why it drives positioning), start with You don't have a positioning problem. You have an ICP problem.
Once your ICP hypothesis is clearer, your homepage has a job: speak directly to that person. For the homepage-side version, read Your website isn't confusing. Your positioning is.
Start with the problem, not the product
Most founders start with their solution and work backwards to find customers. That's the wrong order.
Start with the problem. A specific, observable problem that real people have right now. Not a problem you invented. Not a problem you assume exists. A problem you've seen firsthand — in a job, in a previous company, in your own life.
Then ask: who has this problem most acutely? Who loses the most when this problem goes unsolved? Who already tries to solve it with spreadsheets, workarounds, or overpriced tools that only half-work?
That person is your ICP hypothesis. You haven't validated it yet. But it's a starting point grounded in something real rather than something imagined.
Your own experience is data
If you built something out of personal frustration, you're the first data point.
What's your background? What industry did you work in? What problem did you hit repeatedly that nobody was solving well? The fact that you built this product means you had a reason. That reason points to someone.
You're not the ICP — you're too close to it. But you're a proxy. And a proxy is better than nothing when you have nothing else.
Find the watering holes
Your ICP exists somewhere before they find you. Figure out where.
They're in Slack communities. They're on specific subreddits. They go to certain conferences. They read certain newsletters. They complain about the same things in the same forums.
Go there. Not to pitch. To listen.
Read threads where people describe problems in your space. Note the exact words they use. Note what they've already tried. Note what frustrates them about existing solutions. Note how often the topic comes up and who's engaging with it.
This is qualitative research and it's free. It tells you whether the problem is real, how people talk about it, and whether the people who have it are actually looking for a solution.
Talk to people before you have a product to show
This is the part most founders avoid because it feels uncomfortable.
You don't have a product, so you don't have something to demo. You have a problem hypothesis. Go test it.
Find ten to fifteen people who match your ICP guess. Don't pitch them. Tell them you're doing research. Ask:
- How do you currently handle [the problem]?
- How much time does it cost you? How much money?
- Have you ever tried to fix it? What happened?
- If this problem disappeared tomorrow, what would change?
You're not looking for validation. You're looking for signal — do their answers match your assumptions, or do they reveal something you got wrong?
If multiple people describe the same pain in similar terms, you're onto something. If they look confused or explain why it's not actually a problem, you need to adjust.
Do this before you build. Or if you've already built — do it now. It's not too late.
Borrow credibility from a comparable problem
No customers yet doesn't mean no comparable data exists.
Look at companies who solved an adjacent problem for a similar customer. How did they describe their ICP? Who did they target first? What did their early positioning look like?
This isn't copying. It's orientation. You're looking at how a similar market behaved to build a more informed hypothesis about your own.
Investor case studies, founder interviews, teardowns, early press releases — all of this is public and useful. The patterns in adjacent markets are often surprisingly transferable.
Make a bet and move
At some point the research has to stop and the hypothesis has to be tested in the real world.
Pick one segment. The one that feels most likely based on everything you've learned. Go deep on it. Write copy as if you're talking to that specific person. Reach out to people in that segment directly. See if they respond, engage, convert.
You'll find out quickly whether your hypothesis holds. If it does, you have the beginning of an ICP. If it doesn't, you have data — which is more than you had before.
The mistake is spreading across multiple segments at once because you're not sure. Breadth feels safe. It's not. It just means getting weak signal from every direction and strong signal from none.
What you're looking for
You're not looking for certainty. You're looking for a working hypothesis specific enough to act on.
“Early-stage B2B SaaS founders, two to ten people, who just closed a seed round and are trying to grow from first customers to repeatable revenue — and don't have the budget for a full-time CPO.”
That's a real sentence about a real type of person. It might be wrong. It will probably evolve. But it's specific enough to write copy, do outreach, and make product decisions.
Vague ICPs produce vague results. A wrong specific ICP is easier to fix than a vague one. You test it, you learn, you adjust.
The honest version
Finding your ICP without customers is hard. There's no shortcut that replaces real conversations with real people who have real problems.
What you can do is compress the loop. Make a hypothesis fast, test it with real humans fast, and update it fast. The founders who get this right aren't smarter — they're just more willing to be wrong in public and correct quickly.
Your first ICP will be wrong in some way. That's fine. It just needs to be specific enough to get you to your first customers. They'll tell you the rest.
Once you have a clearer picture of who you're targeting, your homepage needs to speak directly to them. Run the Homepage Teardown to see if it does.
Run the Homepage Teardown → ← Back to blog