A firmographic list tells you who could buy. It has nothing to say about who will. Here is what I look for instead, and the trap that catches founders with early traction.
You built the list properly. Four hundred companies, every one a match on size, vertical and stack. You wrote a sequence you were happy with. Three weeks later you have nine replies, two of them polite versions of no, and no way to tell which part to change.
So you rewrite the email. Then the subject line. Then you try a different sequence tool.
I have watched founders spend an entire quarter in that loop, and the email was almost never the thing that was wrong.
Your filter can only see what a company is
Here is the part that took me a while to say clearly.
Firmographics describe what a company is. Headcount, industry, funding stage, tech stack. All of it stable. All of it still true six months from now.
Buying is set off by what just changed inside a company.
Take two accounts that match your filter line for line. Same size, same vertical, same CRM. One of them hired a VP six weeks ago who owns this number and needs something to show by the end of the quarter. The other has run the same process for three years, with next year's budget already committed elsewhere.
Your filter sees one account. Your reply rate sees two.
Why that gets expensive
The cost is not the low reply rate. It is that you cannot diagnose it.
When your list is built on fit alone, a quiet week tells you nothing. The message may have been wrong. The timing may have been wrong. The account may never have been going to move at all, in which case the best email ever written would also have gone unanswered.
So you change the thing that is easiest to change, which is the copy. And the real problem survives another month.
I have had this conversation enough times to recognise it from the first sentence. The founder describes a messaging problem. Twenty minutes into looking at the actual accounts, it turns out the list had no reason to reply.
The question I add to the four boxes
When I open an ICP document, I am looking for one thing that is usually missing:
What has to be going on inside a company for this to become urgent?
Everything else I work on follows from that question. It is the difference between a list of companies that could plausibly use your product and a list of companies with a reason to answer you this month.
I call the result a situational ICP. It has five parts rather than one.
Fit
The firmographics you already have. Keep them, and keep them loose.
Fit is a boundary, and its job is to clear away the obviously wrong so the other four have room to work. If your fit criteria return fewer than a few hundred companies, they are doing work that belongs further down this list.
The job
What is your buyer accountable for getting done? Not the feature they asked about in the demo. The outcome someone will ask them about in a review.
A head of support evaluating a help desk is rarely buying a help desk. They are buying a way to close more tickets without more headcount, because the headcount request came back declined in March.
You will find this in job descriptions for the role, in how people describe the problem in community threads, and in the first ninety seconds of any sales call where you ask what made them look at this now and then say nothing.
The situation
This is the one that changes your results, and the one most ICP documents skip entirely.
A situation is a change inside the account that makes your problem expensive this quarter instead of next year. The ones I see repeat across B2B SaaS:
A new leader taking over the function
A funding round, which turns a nice-to-have into a line in a plan
A compliance or audit deadline with an actual date on it
A migration or replatform already underway
Sudden hiring in the team that feels the pain
A competitor announcement that resets what the board expects
The test for a good situation is that it has a clock attached. Something moved, and the cost of doing nothing about it goes up every week.
The champion
Titles tell you where someone sits. What you need to know is what they are measured on, because that determines whether they will spend political capital on you.
Two people with the same title at two companies are often completely different buyers. One is measured on cost per ticket. One is measured on time to resolution. Your first sentence to each of them should be different.
The current way
Your buyer already solves this problem somehow. A spreadsheet, a junior hire, a competitor, a process that mostly works.
That is your real competition, and the gap between what it does and what they now need is your positioning. Write down what they use, what it handles adequately, and the exact point where it stops working. That last part is the sentence your homepage should open with.
If you already have customers, read this part twice
Everything above assumes you are working out who to go after. If you already have five or six customers, you might reasonably think you can skip it and just describe the people who bought.
That is the trap I see most often, and it is worth slowing down for.
What you have is an early customer profile. It is real data, drawn from real money, and it is genuinely valuable. It is also answering a narrower question than it looks like it is answering: who said yes to you, so far.
An ideal customer profile answers a different question: who is this worth the most to, whether or not there is a relationship in the room.
The two documents look identical. Same fields, same shape, same confidence. The gap between them comes from how those first customers arrived.
The trait that hides inside your early profile
If your first eight customers came through your own network, then one of the things they have in common is that they knew you, or they knew someone who did.
That trait was present in every one of those deals. It influenced all of them. And it appears in no column of any document you will ever write.
I see the same pattern regularly. A founder lines up their eight customers, notices they are all 30 to 50 person agencies in the UK, and writes that down as the ICP. Six of the eight came through one former employer's alumni network. The firmographics were a coincidence riding along with the introduction, and the next two quarters of outbound are built on it.
A five-minute test you can run today
Open your customer list. For each one, write a single line: how did the first conversation start?
Inbound, they found usCold outreach, no prior contactWarm intro from someone we knowMet at an eventExisting relationship from a previous company
Now group them.
If more than half begin with a relationship, your profile is describing your network as much as your market.
That is a normal place to be. It is how most good companies start, mine included. It just means the profile needs a second source of evidence before you build a motion on top of it.
What your early customers are genuinely good for
Quite a lot, as long as you ask them the right question.
They are excellent on what changed after they started using it. That is first-hand, specific, and it is the raw material for your positioning. Ask for it in their words and use their words.
They are less reliable on what made them look in the first place, because part of the answer was that they trusted you. The trigger is the thing you most need, and a warm relationship is exactly the thing that hides it.
So hold what they tell you against what the wider market says, rather than treating it as the whole answer.
Where your evidence should come from
The right source depends on where you are, and choosing wrong here is the most expensive mistake in the whole exercise.
Where you are | Your primary evidence | What your customers add |
Product built, no customers yet | The market: reviews, communities, job posts, industry discussion | Nothing yet, and that is fine |
Five or six, mostly through your network | The market first, cross-checked against your customers | What changed after they started |
A real customer base won without introductions | Your own customers, read as one set | Almost everything you need |
The middle row is the one that catches people out. Having customers feels like having answers. When those customers arrived through trust rather than through pain, the market work still has to happen. Your customers make it faster and sharper.
Make it something you can act on
An ICP you cannot act on is a research document. The working test for each of the five parts is whether somebody outside the company can see it.
Part | Visible from outside? | Where you find it |
Fit | Yes | Any data provider |
The job | Often | Job posts, product pages, public roadmaps |
The situation | Sometimes, and this is the valuable part | Leadership changes, funding news, hiring spikes, compliance dates |
The champion | Yes | Title, tenure, what they post about |
The current way | Often | Tech stack data, review sites, community threads |
The parts you can observe become your targeting. The parts you cannot become the questions you ask on the first call. Both are useful. Only one of them builds a list.
Write it on one page
If it runs longer, you have written research rather than an ICP.
FIT The company profile that could plausibly use thisTHE JOB What they are accountable for getting doneTHE SITUATION The events that make it urgent, and how we spot each oneTHE CHAMPION Their role, and what they are measured onTHE CURRENT WAY What they use today, and where it stops workingEVIDENCE Where each line came from, and how confident we are in it
That last block does more work than it looks like it does. Six months from now somebody will ask why the ICP says what it says. If the answer is written down, the document can be argued with and improved. If it is not, it quietly turns into folklore, and nobody will touch it.
Where to start, depending on where you are
Which part of this to do first depends on what you already have.
If you have built the product and have no customers yet, the whole profile comes out of market evidence, and that is exactly the work in Find the Demand.
If you have five or six customers and most came through people you know, the job is to read the market first and hold your customers against it, so the early profile becomes an ideal one. That is Find the Pattern.
If you have a real customer base and a team telling slightly different versions of the story, your own customers are the strongest evidence you have, and the work is settling one account of who this is for. That is Build the Motion.
If you would rather work out which of those three you are before reading further, Diagnose My GTM is a short set of questions that will tell you.