How to Define Your Ideal Customer Profile (ICP)

Most lead lists die the same way: someone exports five hundred companies, sends the same message to all of them, and gets three replies — two of which say "remove me." The list was never the problem. The targeting was. If you cannot state precisely who you sell to, no volume of outreach will save you.

Treat this article as a working session, not theory. Set aside forty minutes, open a blank document, and answer the questions in each section as you read. By the end you will have a one-page ideal customer profile (ICP) you can hand to a cofounder, a freelancer or a lead-generation tool — and get back a list that is actually worth contacting.

What an ICP is — and what it is not

An ideal customer profile describes the company most likely to buy from you, pay you well, stay with you and refer others. It is a firmographic definition: industry, location, size and observable signals. It is not a fictional persona with a name, an age and a favourite podcast.

Keep the two apart. A buyer persona describes the person you talk to — the owner, the marketing manager. The ICP describes the company you target. You need the ICP first, because it decides which doors you knock on; the persona only decides what you say once a door opens.

One rule before we start: an ICP is a hypothesis built on evidence, not a wish list. Every answer below should come from your real customers — the deals that closed fastest, at the best margins, with the fewest headaches. If you are pre-revenue, write your best-informed guess and mark it clearly as a guess to be tested.

Step 1: Niche — the vertical you actually win in

Start from deals you have already won. List your last ten customers and answer four questions:

• Which industry closed fastest, with the fewest objections?
• Who paid your price without negotiating it down?
• Who came back for more, or referred someone else?
• Where do you have a concrete result you can retell as a story?

The pattern that emerges is your niche — and it is usually narrower than what you say at networking events. "We serve small businesses" is not a niche. "Dental clinics" is. "Dental clinics that run paid ads" is even better, because it adds a behaviour you can observe from outside.

Write the first line of your one-pager: Niche: one vertical, in plain words. If you are torn between two, write both but rank them. Your first campaigns should hit the top one only; the runner-up becomes your comparison group later.

Step 2: Geography — where you can deliver and sell

Founders skip geography because "we can work with anyone remotely." Perhaps — but sales has a geography even when delivery does not:

• Language. Outreach in the prospect's native language reliably outperforms English-to-everyone.
• Time zones. Calls across nine hours quietly kill deals that a same-day reply would have closed.
• Local proof. A case study from one city closes deals in that city. Social proof travels badly.
• Rules and payments. Invoicing, taxes and compliance differ enough by country to change your margins.

Choose a primary territory you can name precisely: a city, a metro area, one country. "Berlin and Hamburg" beats "the DACH region" for a first campaign — you can widen later, once the narrow version works. Write it down: Geography: named places, not regions.

Step 3: Company size — the band that buys at your price

Size is a proxy for budget, decision speed and who sits across the table. Employee headcount is the easiest version to observe from outside; use revenue only where it is published. The bands behave differently:

• Solo and micro (1–9). Decide in a day, but budgets are thin and churn is high.
• Small (10–50). The sweet spot for many services: real budget, and the owner still answers messages.
• Mid-size (50–200). Bigger contracts, but you now sell to a group, and cycles stretch to months.
• Enterprise (200+). Procurement, security reviews, pilots. Only go here if your cash flow survives six-month cycles.

Pick a band, not a point — wide enough to fill a list, narrow enough that one message fits everyone in it. Write it down: Size: for example, 5–50 employees, one location or a small chain.

Step 4: Budget signals — public evidence they can pay

You cannot read a company's bank statement, but you can read the traces money leaves in public. Budget signals are observable facts that correlate with the ability — and habit — of paying for things:

• They run paid ads (search their name in ad libraries, or just watch their Instagram).
• They are hiring — job listings are spent money before the hire even starts.
• Their website is professional and was updated recently, not abandoned since 2019.
• They have several locations, or their own prices sit above the local median.
• They already pay for tools or services adjacent to yours.

Pick two or three signals you can verify in under a minute per company. Then define the reverse — signs the budget is not there: dead social accounts, a website untouched for years, no reviews anywhere. Write it down: Budget signals: what you check, and what disqualifies.

Step 5: Negative ICP — who you refuse to contact

This is the most underrated section of the workshop. A negative ICP is a written list of disqualifiers — companies you will not contact even when they look reachable. Bad-fit customers are not neutral: they consume your pipeline with long cycles, discount pressure, painful delivery and early churn.

Mine your history again, this time for regret:

• Which deals dragged for months and then died?
• Which clients churned within a quarter?
• Who cost more to serve than they ever paid?
• Which industries created legal or reputational friction for you?

Typical entries: companies below a size floor, marketplaces and aggregators instead of actual operators, franchises where decisions happen at headquarters, verticals you cannot reference publicly. Write it down: Negative ICP: "We do not contact…" — and mean it when the list is thin and temptation is high.

Assemble the one-pager

You now have five lines. Put them on a single page, in this order, and add two lines that make the profile usable in outreach:

1. Niche: the vertical you win in.
2. Geography: named cities or one country.
3. Size: the employee band.
4. Budget signals: two or three checks, plus disqualifying signs.
5. Negative ICP: who you refuse to contact.
6. Trigger: the event that makes them buy now — a new location, a new hire, a seasonal peak.
7. Proof: the one result you cite in the first message.

A filled-in example: Niche — dental and aesthetic clinics. Geography — Lisbon and Porto. Size — 5 to 30 staff, one to three locations. Budget signals — running Instagram ads, prices above the city median, hiring front-desk staff. Negative ICP — solo practitioners, franchise chains, clinics without a website. Trigger — opening a second location. Proof — 40% more bookings in 90 days for a clinic in Braga.

Pressure-test it against a real list

An ICP that lives only in a document is still a guess. The test is mechanical: turn the profile into an actual list of companies and look at them one by one.

Pull fifty companies that match your niche and geography. Then score them against the rest of the page: how many fall inside your size band? Can you actually see your budget signals from public information — the website, the social profiles, the contact channels they use? How many trip a negative-ICP filter you forgot to write down?

This is the loop where a search tool earns its keep. You can build your first ICP test list with JustLeadIt — two searches are free: it finds companies by niche and city across maps, business registries and web search, and collects their public contacts — email, phone, WhatsApp, Instagram, Facebook, LinkedIn, website — so scoring fifty companies against your one-pager takes an evening instead of a week. Export the scored list to XLSX or CSV and mark fits and misses right in the file.

Read the results like a practitioner. If fewer than half the companies look like plausible fits, your niche or geography is too broad — tighten one of them, not both at once. If nearly everything fits but you cannot detect a single budget signal from outside, your signals are not observable — replace them with ones that are. Each revision takes minutes, and each one compounds.

Revisit it every quarter

Your ICP is versioned, like software. After every hundred contacts — or every quarter, whichever comes first — reopen the one-pager with three questions: which segment replied most, which actually closed, and which churned or became painful. Promote the real winners into the profile; demote the regrets into the negative ICP. The one-pager is not a poster on the wall. It is the operating filter for every list you build next — and the teams that revise it quarterly are the ones whose reply rates climb instead of decay.