ICP Strategy

How to define your ICP for B2B SaaS
whether you have customers or not

Pre-revenue or $300K ARR — the ICP problem looks different at each stage but the cost of getting it wrong is the same. This article addresses both, because the fix is different depending on where you are.

Amulya S Kashyap
Founder, Wiremap  ·  20 years B2B GTM
June 2025
10 min read
Jump to your stage

Most ICP documents are useless regardless of stage. They sit in a Notion page, get updated once a quarter by someone who does not run outreach, and influence exactly nothing. The pre-revenue founder's ICP is useless because it is entirely hypothetical. The growth stage founder's ICP is useless because it describes who they want to sell to rather than who is actually buying. Both need fixing — but in completely different ways.

I am going to address both stages in this article because the ICP problem at each one is genuinely different. If you are pre-revenue, you have no closed-won data and the most dangerous thing you can do is commit too early to an ICP that is based entirely on assumption. If you are at growth stage with some paying customers, the most dangerous thing you can do is ignore the pattern in what has already happened and keep building on an ICP that does not match reality.

The universal principle that applies to both: your ICP is not who could benefit from your product. It is who has the urgency, the authority, and the budget to buy this quarter. That is a much smaller and more specific group — and finding it is the highest-value commercial exercise you will do at any stage.

First — why most ICP documents produce nothing

Before splitting by stage, let me explain why the typical ICP definition fails regardless of how carefully it was written.

A market description — not an ICP
"B2B SaaS companies with 50 to 500 employees that need to improve team productivity and operational efficiency."

This could describe 200,000 companies. There is no named buyer. No trigger event. No reason why this company would act this week rather than next quarter.
An actual ICP — buildable into a list
"Heads of Operations at B2B SaaS companies between 30 and 150 employees that raised a seed or Series A in the last 12 months and are scaling headcount for the first time."

You can build this list in LinkedIn Sales Navigator this afternoon. The trigger — recent raise, first scaling moment — creates timing and urgency.

The test of a real ICP is simple: can you build a list of 500 companies from it right now using four filters in LinkedIn Sales Navigator? If the answer is no, you have a market description. The work — whether you are pre-revenue or at $300K ARR — is to narrow it until the answer is yes.

The three layers every ICP needs regardless of stage

LayerWhat it isWhy it matters
Layer 1
Firmographics
Company size, industry, geography, funding stage, tech stack. The shape of the company. Determines who goes on the list. Necessary but not sufficient on its own — 5,000 companies that fit the firmographic profile is a universe, not a target.
Layer 2
Buyer profile
The specific job title of the person who owns the problem and has budget authority. What they are measured on. Who else is in the buying decision. Determines who to contact at the company. A Head of Operations and a VP Engineering at the same company have completely different priorities and the message that converts one will not convert the other.
Layer 3
Trigger events
The specific event in the buyer's world that creates urgency to act now rather than defer the decision. A funding round, a team change, a failed process, a product launch, a growth threshold crossed. Determines when to reach out and what the first sentence of every message says. This is the layer most founders skip. Without it, even the right buyer at the right company has no reason to act this week.

Most ICP documents only contain Layer 1. Some add a rough version of Layer 2. Almost none name specific trigger events. Which is why most outreach messages read like vendor pitches rather than conversations that start with the buyer's current situation.

Pre-revenue & Early Stage
Building your ICP when you have no customers yet

This is the hardest version of the ICP problem because you have no data to validate against. You are working entirely from hypothesis — and hypothesis is the least reliable source of ICP intelligence that exists. Every founder has a bias toward the buyer they wish would buy. The product was probably built with a specific buyer in mind. That buyer may or may not actually convert when you reach them.

The pre-revenue trap: Committing hard to an ICP before you have tested it produces an entire outbound motion built on an untested assumption. The most expensive version of this is hiring an SDR, building sequences, and running six months of outreach before discovering that the buyer you targeted does not have the urgency, the budget, or the authority you assumed they had.

How to build a hypothesis ICP that you can actually test

At pre-revenue, the goal is not a perfect ICP. The goal is a testable hypothesis that is specific enough to run an experiment against. Here is the framework.

1
Start with the problem, not the product
Name the specific problem your product solves in one sentence. Not the product category — the problem. "Teams waste two hours per day switching between tools to track project status" is a problem. "Productivity software" is a category. The problem statement points you toward the buyer who has that problem acutely enough to pay to fix it.
2
Name who has that problem most acutely right now
Not who could benefit broadly — who has it so badly that it is already costing them something measurable. A team of 8 people who just went remote and are losing visibility across projects has this problem more acutely than a team of 80 with existing process. Acute pain, not theoretical benefit, is what drives buying decisions.
3
Identify a trigger event that makes this urgent right now
What just happened in their world that makes this week the right moment? For a productivity tool targeting remote teams: a company that switched from office to hybrid in the last 3 months has just discovered the problem. For an AI-driven risk tool: a company that just raised a Series A and is scaling its first operational team suddenly needs process where none existed before. The trigger is what creates timing — and timing is what converts interest into a paid conversation.
4
Build a list of 50 and run the experiment
Not 500. Fifty. Build a list of 50 companies that match the hypothesis ICP — the firmographic profile, the buyer title, and the trigger event. Run 50 personalised outreach messages. Track reply rate, conversation quality, and whether the buyer recognises the problem you are describing. The data from those 50 conversations is worth more than any amount of ICP theorising.

What a successful experiment looks like: A reply rate above 8% from a tightly defined 50-person list, with at least 5 conversations where the buyer says something like "yes, we have exactly that problem" without you having to explain it. That is validation. Below 5% reply rate means the hypothesis is wrong somewhere — either the buyer title, the trigger event, or the problem framing.

Anonymised client pattern
AI-driven workflow tool — Pre-revenue — UK and US
Early Stage

The founder had built an AI tool that automated project status updates across Slack, Jira, and Notion — eliminating the manual check-in process that ate 90 minutes of every project manager's day. The initial ICP on paper: "product and engineering teams at B2B SaaS companies."

That described approximately 300,000 teams globally. The outreach reply rate was 2.3% over two months. The conversations that did happen were polite but not urgent — people agreed the problem existed but were not moving toward paying to fix it.

The diagnosis surfaced a more specific pattern: the pain was most acute at companies between 20 and 60 people that had just adopted Jira for the first time after growing beyond spreadsheets. At that specific transition point, the status update problem becomes acute enough to pay for immediately. The outreach was rebuilt around that trigger — companies that had added Jira to their tech stack in the last 6 months — and reply rate moved to 11.4% in the first four weeks.

Result: ICP narrowed from 300,000 teams to 4,000 specific companies. Reply rate from 2.3% to 11.4%. Same product. Different trigger.

The aspiration question — and when it helps

Every early stage founder has an aspirational ICP — the enterprise customer, the global brand, the category leader. Aspiration is not useless. It is useful in one specific way: it tells you the direction you are building toward. It is not useful as a starting point for outreach, because enterprise buying cycles, procurement requirements, and social proof expectations are completely different from early adopter buying behaviour.

The framework that works: start with the buyer who has the most acute pain and the least friction to buy — typically a company small enough to make fast decisions, large enough to have budget for the problem. Build your first 10 customer relationships there. Use those outcomes to create the social proof that makes the aspirational buyer reachable in 18 to 24 months. Aspiration is a destination, not a starting point.

Growth Stage
Fixing your ICP when you already have customers

The growth stage ICP problem is the opposite of the pre-revenue problem. You have data. You have paying customers. You have 12 to 18 months of outreach history and pipeline data. The problem is that most founders have never interrogated that data to understand what the actual pattern is — and they are still running outreach based on the hypothesis they had before any of that data existed.

The single most common finding in Wiremap growth stage diagnoses: the written ICP and the actual closed-won profile are meaningfully different. The written ICP says one thing. The CRM data says another. The outbound motion is targeting the written ICP. The deals that are closing are coming from somewhere else.

How to extract your real ICP from closed-won data

1
Pull your last 10 closed-won customers — right now, not later
For each: company name, company size, industry, geography, the exact job title of the person who signed the contract, what triggered the conversation, how long the sales cycle was, and deal size. Put it in a spreadsheet. Do not try to find the pattern yet — gather the data first.
2
Look for the gap between your written ICP and what actually bought
This is where the most valuable insight lives. You may have been targeting VPs but your actual buyers were Heads of Operations. You may have been targeting companies with 200 employees but your deals kept closing at companies with 40. You may have been targeting one industry but three of your last five deals came from an adjacent one you had not focused on. The pattern in the data is your real ICP.
3
Identify the trigger event that preceded each purchase
For each closed-won customer, ask: what was happening in their world in the 30 to 90 days before they bought? This is the question most founders have never asked their own customers. Call three of them. Ask directly: "When did you first realise you needed to solve this problem and what had just happened?" The trigger is almost never "we were generally looking for tools like yours." It is almost always a specific event — a team growth threshold, a failed audit, a new hire who changed priorities, a product launch that exposed a gap.
4
Rewrite your ICP from what the data shows — not what you assumed
If your last seven customers were Heads of Operations at Series A SaaS companies between 30 and 80 employees that had just hired their first Head of Sales, your ICP is that. Not "early stage SaaS companies that need operational tooling." The former produces a list. The latter produces a description. The moment you commit to the data-derived ICP, every outreach message, every sales conversation, and every channel decision becomes sharper.

The growth stage trap: The most expensive version of the growth stage ICP problem is scaling an outbound motion before interrogating the closed-won data. Hiring SDRs, building sequences, engaging a demand gen agency — all of it pointed at an ICP that does not match the pattern of who actually buys. The investment is real. The return is limited because the direction is wrong.

Anonymised client pattern
AI risk management platform — $180K ARR — US mid-market
Growth Stage

The company had built an AI-powered operational risk monitoring tool for mid-market businesses. Written ICP: "Operations leaders at mid-market companies in regulated industries." The outbound motion had been running for six months targeting Operations Directors and VP Operations at financial services and healthcare companies with 200 to 1,000 employees. Pipeline was thin and sales cycles were long.

A closed-won analysis of their 11 customers revealed a different pattern entirely. Eight of the eleven were Head of Risk or Chief Risk Officer at technology companies — not traditional regulated industries — between 80 and 250 employees, and seven of those had raised a Series B within the previous 18 months. The trigger event in almost every case was an investor or board requirement to formalise risk management as part of post-round operational maturity.

The outbound had been going to the wrong industry, the wrong title, and missing the trigger entirely. The product was excellent. The direction was wrong. The ICP rewrite pointed the same product at a more specific buyer with a documented trigger event — Series B SaaS companies under new board pressure to formalise risk function.

Result: 11 customers analysed. ICP completely rewritten from closed-won data. Outbound rebuilt around Series B trigger. First qualified pipeline from new ICP in 3 weeks.

What to do about aspirational buyers at growth stage

At growth stage, aspiration is more dangerous than at pre-revenue — because you have enough momentum to convince yourself the enterprise deal is possible, which leads to long sales cycles that produce nothing while the core ICP motion stalls.

The framework: run the core ICP motion for 80% of your outbound effort. Reserve 20% for one or two aspirational targets where you have a genuine warm introduction or a specific reason to believe the timing is right. Do not build your pipeline forecast around aspirational deals. Do not resource your outbound team primarily against the ICP you want — resource it against the ICP that converts.

Both stages
The universal ICP test

Run this on your current ICP — regardless of stage
Can you answer yes to all five of these?
  • I could build a list of 500 companies that match this ICP right now using LinkedIn Sales Navigator with four or fewer filters
  • I know the specific job title of the person I am reaching out to — not "senior leadership" or "decision makers" but a named title
  • I can name a specific external trigger event that creates urgency for this buyer right now — something in their world, not in mine
  • I could write the first sentence of an outreach message that would make this buyer think "how did they know" because it names their exact current situation
  • I could describe a deal that does not fit this ICP and explain exactly why — meaning the boundaries are clear enough to disqualify as well as qualify

If you answered no to any of these, your ICP needs narrowing before any outreach changes will produce meaningfully different results. The number of no answers tells you how much work is needed.

The single thing that compounds everything else

A precise ICP does not just improve outreach. It makes every other part of GTM more effective simultaneously. The message converts better because it is specific. The channel produces better results because you are reaching the right buyer. The sales conversation moves faster because the urgency is real. The close rate improves because the product actually solves the problem the buyer has right now.

None of this happens at the market description level. It all starts with the buyer-level specificity that most ICP documents never reach — regardless of how many hours were spent writing them.

At pre-revenue: the specificity comes from a hypothesis built around a trigger event, tested quickly against 50 real people. At growth stage: it comes from interrogating your closed-won data and letting what has actually happened tell you what the real ICP is.

Both paths lead to the same place. A list you can build this afternoon. A message that names the buyer's situation before it describes your product. Outreach that converts because it was written for one specific person — not for everyone who could theoretically benefit.

Two products. One for each stage.

Not sure if your ICP is specific enough?

The GTM Clarity Assessment is built for pre-revenue and early stage founders who need to validate their ICP hypothesis before building a motion on top of it. The GTM Risk Assessment is built for growth stage founders who have data but have not interrogated it properly yet.

No calls required. Intake form sent within 30 minutes of payment. Delivered in 7 business days.