A founder I talked to last quarter spent four months running customer discovery interviews. Forty-two calls. A Notion database with color-coded tags. At the end of it she sent me a one-page document describing her ideal customer profile b2b target: regional logistics companies with 80–300 trucks, an ops director who’d been promoted into the seat in the last 18 months, currently running a paper-based dispatch process they were patching with spreadsheets. I read it and asked when she’d first known. She paused. “Probably the second week I worked at my last company.”
This is the pattern. Operators run formal discovery processes to arrive at conclusions they could have written on day one. The process isn’t wrong — it just isn’t what they think it is. They aren’t discovering the customer. They’re translating ten years of pattern recognition into something they can hand to an investor or a salesperson.
The faster you skip to that step, the faster you build something the buyer will pay for.
What an ideal customer profile actually is
An ICP is operational, not aspirational. It describes the company characteristics that make a buyer worth pursuing — not a portrait of one human. Personas live downstream. The ICP comes first because it tells you which companies even get a phone call.
The mistake most early-stage founders make is writing an ICP that reads like a marketing segment. “Mid-market SaaS companies with growing sales teams” isn’t an ICP. It’s a category. An ICP names a company tight enough that you could open LinkedIn right now and list twenty real companies that fit. If you can’t, the profile is too loose and your go-to-market motion will burn cash chasing buyers who were never going to convert.
What belongs in a B2B ICP
A useful profile names five things:
- Company size by revenue range, not headcount. Headcount lies. A 200-person services firm and a 200-person software company have nothing in common as buyers.
- Specific vertical. “Professional services” is not a vertical. “Mid-market accounting firms with audit practices” is.
- The pain trigger. The event or condition that surfaces the need. New compliance regime. A bad quarter. A senior hire who wants to fix the workflow they inherited.
- Budget owner identity. Title, function, and seniority of the person who can actually sign. Not the user. The signer.
- The maturity floor. Below what size, what process sophistication, or what revenue level, the buyer will not show up no matter how good the pitch is.
If you can’t fill those five lines from memory, you haven’t compressed your experience into a profile yet. Doing so is most of the work.
Why operators have an ICP advantage
Operator founders are working with a dataset no consultant can reproduce. Ten years inside an industry produces thousands of micro-observations about which companies buy, which cancel, which stall in procurement for six months and never close. The information is already in the founder’s head. It’s not labeled, but it’s there.
The reason a research process feels necessary is that the data hasn’t been translated into language yet. Talking to forty buyers doesn’t teach the operator new information — it forces them to articulate what they already know in a form someone else can read. That’s worth doing, but it can be done in three weeks with eight conversations, not four months with forty.
The faster path: write the ICP from memory first. Then run a small number of targeted calls to test the parts you’re least sure about. Most operators are surprised to find that the parts they were least sure about were the right parts to test, and the parts they were certain about hold up.
How to sharpen the ICP you already have
Three questions surface most of what you need:
Who was the most successful early adopter of similar tech in your last decade, and what was true about them? Not the logo. The conditions inside the company when the deal closed. Who was new in the seat. What problem had just blown up. What budget had just been approved. The shape of that company is closer to your ICP than any spreadsheet of segment data.
Who was the disaster customer and why? The deal that closed but never produced value. The implementation that failed. The pilot that went six months past its deadline. Disaster customers tell you what’s below the maturity floor — the conditions that make the product structurally unable to land, no matter how strong the relationship.
What did the serious buyer say in the first meeting that told you they were serious? There’s almost always a specific phrase, a specific question, or a specific document the serious ones produce that the tire-kickers don’t. Whatever that signal is, it belongs in your ICP as a qualification criterion.
Write the answers. Look at the pattern. That’s your ICP.
What to do with it
Write it before your first outbound call. Not because the first version will be right — it won’t — but because you need a target to update. Founders who start outbound without a written ICP end up calling everyone, learning nothing in particular, and a quarter later have a contacts list instead of a profile.
Update it after every disqualified call. The reason a buyer wasn’t a fit is data. If you keep finding companies who looked right on paper and bounced for the same reason, that reason belongs in the ICP — as either a refinement of who fits or a new line in who doesn’t.
Hand it to your seed investors before they ask. Investors don’t expect the ICP to be permanent. They expect the founder to have a sharp opinion they’re updating against evidence. An operator who can name the exact buyer type — size, vertical, trigger, signer, floor — is immediately more credible than a founder describing “the broader market.” The first founder sounds like they’ve sold to this person. The second founder sounds like they’re still hoping to meet them.
And when you build for vertical SaaS specifically, the ICP is the product spec. Every feature decision routes back through it. If the maturity floor is “companies with at least three dispatchers,” then no feature should be designed for a one-dispatcher company. The ICP isn’t a marketing artifact. It’s the constraint that keeps the company from drifting.
You already know the answer. Write it down today.