The ICP Worksheet

The cost of not having one

When your ICP is “B2B companies that need what we do”, you are effectively targeting everyone, which means you target no one well. Outreach goes out wide and cold, reply rates stay low, and the deals you do land look nothing like each other, so there is no pattern to build on.

A sharp ICP does the opposite. It tells you exactly which companies to approach and which to leave alone, so every hour of outreach lands on someone who could realistically buy. It is the single biggest lever on whether lead generation works at all.

What good looks like

This is the standard. A useful ICP is not a persona, it is a set of filters precise enough to build a list from. Measure yours against it. A strong one has:

• Firmographics you can filter on. Industry, company size, location and business model, in terms a list-building tool understands.UK-based B2B agencies, 10 to 50 staff, £1m to £5m revenue.
• A trigger, not just a profile. The signal that a company is likely in-market now, rather than simply a fit on paper.Recently hired a sales role, raised funding, or opened a new location.
• The role you sell to. The job title of the person who feels the problem, and the one who signs the cheque. Often not the same person.Pain felt by the Head of Sales, signed off by the Founder or MD.
• What makes a great fit, not just an acceptable one. The two or three attributes your best customers share that the merely tolerable ones do not.Already running outbound, but inconsistently and without results.
• Clear disqualifiers. Who to skip on purpose, even when they could technically buy. This protects your time as much as the targeting does.Under 10 staff, pre-revenue, or purely consumer-facing.

The test: you could hand this to a stranger and they could build a list of 200 real companies that match it, without asking you a single question.

What it looks like in practice

Here is a worked ICP for an agency that sells outbound lead generation. Yours will differ, but this is the level of specificity to aim for.

Three questions to sharpen your own:

• Which of your last ten customers were the easiest to win and the best to serve? What do they share?
• Which deals did you regret taking on? What did they have in common?
• What has to be happening inside a company for them to need you now, rather than next year?

How to build it yourself

1. 1
   Start from your best customers, not your hopes. List your five or ten best accounts: easy to win, profitable, still around. They are the evidence of who you should be chasing.
2. 2
   Find what they share. Look for the firmographics in common: industry, size, location, business model. Patterns you can filter on are what you are after, not vague traits like “ambitious”.
3. 3
   Name the trigger. Work out what tends to be true when a good-fit company actually buys. A new hire, a funding round, a leadership change. This is what turns a static list into a timely one.
4. 4
   Pin down the buyer. Identify the title that feels the problem and the title that signs it off. Your outreach has to speak to the first and be acceptable to the second.
5. 5
   Write the disqualifiers. Be explicit about who to skip. A clear no protects your time and stops the list filling with companies you will only regret winning.
6. 6
   Test it on a real list. Try to build a list of 200 companies that match. If you cannot, or the matches are clearly wrong, the definition is still too loose. Tighten it and try again.

Build it faster with AI

Short on time? Paste this into ChatGPT, Claude, or any AI assistant. It will interview you and produce a first draft. Treat it as a draft, and only that. A model cannot weigh the nuances of how your business actually sells, and it cannot install the process into your CRM or run it for you. But it is a fast way to a solid first version.

You are a B2B go-to-market expert. Help me define a sharp ICP I can build a target list from, not a vague persona.

First, ask me these one at a time, and wait for each answer:
1. What do you sell, and what problem does it solve?
2. Who are your five best customers, the ones that were easy to win, profitable, and stuck around? What do they have in common?
3. Who has been a bad fit: hard to win, painful to serve, or quick to leave?
4. What has to be true about a company for them to need what you sell right now?

Then, using my answers, produce:
- A firmographic profile I can filter on: industry, company size, location, business model
- The trigger events that suggest a company is in-market now
- The job titles of the person who feels the problem and the person who signs it off
- The two or three attributes that separate a great-fit customer from a merely acceptable one
- A clear list of disqualifiers: who to skip even if they could technically buy

Finally, sanity-check it: could someone build a list of 200 real companies from this without asking me anything? If not, tell me exactly what is still too vague.

Keep it specific to my business. Do not pad it.