How to Choose Your First Ideal Customer Profile
Founders before or just after launch should choose a first ideal customer profile by comparing two or three candidate segments on the same evidence: a buying trigger visible from outside, a named buyer role with spending authority, money already spent on the problem and a product that can be delivered without custom work. Pick the segment with the most observed evidence, not the biggest market. The main constraint is separating what you saw from what you assume.
The choice and its business consequences
This guide is for founders before or just after launch who have two or three plausible customer segments and need to commit to one before spending on sales, marketing or product work. It covers how to compare the candidates on evidence and make a decision you can revisit. If you have not yet talked to buyers in any of the segments, start with how to run a customer interview without pitching and come back with notes.
An ideal customer profile, or ICP, describes the type of account you will sell to first: the kind of organization, the event that makes it act, the person who approves the purchase and the budget the money comes from. Choosing one is a resource decision. It sets which accounts the founders contact each week, what the first product version has to do, which proof points the pitch needs and what the first case study will look like. A wrong choice is expensive mainly in time: months of conversations with buyers who agree the problem is real but have no reason to act now.
The common error is choosing the segment with the biggest apparent market. A large segment that cannot be reached, has no clear trigger or has no budget owner produces slow, inconclusive learning. A narrow segment where the buying moment is visible and the buyer is named produces fast answers, even if it looks small on a slide. The stage-by-stage go-to-market guide places this choice at the start of pre-launch work. This article covers how to make it.
Observable buying trigger
A buying trigger is an event that turns a tolerated problem into an urgent one: a new regulation taking effect, a leadership change, a funding round, a hiring push, a failed audit, a seasonal peak or a weather event. The question for an ICP decision is whether you can see the trigger from outside the account. If you can, you can time outreach to the moment the problem is expensive. If you can only learn about it by asking, every account costs a conversation before you know whether it is in market.
Grade the trigger evidence for each candidate segment:
- Observed. You can name accounts where the trigger happened recently, from a public source such as job postings, permit filings, press releases or local news, and buyers confirmed that it made the problem urgent.
- Reported. Buyers describe a trigger in interviews, but you cannot detect it from outside.
- Assumed. The trigger sounds plausible, but no buyer has described it.
A segment whose trigger is only assumed is not ready to be a first ICP. A segment with a reported trigger can still work if the trigger recurs often enough that a steady share of accounts is in market at any time.
Specific buyer role
The ICP needs a job title, not a department. Name the person who feels the problem, the person who approves spending and, if different, the person who can block the purchase. In small companies these are often the same person, which is one reason first ICPs tend to work better in smaller accounts. In larger accounts the user may be enthusiastic while the purchase is decided elsewhere, through a procurement process the founder cannot see.
Evidence for the buyer role is a record of how the last related purchase happened: who suggested it, who compared options and who signed. A buyer who describes the approval path for a recent purchase is stronger evidence than one who says they would champion a new tool. Before committing, check that the role exists in meaningful numbers in the target industry. The Bureau of Labor Statistics Occupational Employment and Wage Statistics program produces employment and wage estimates annually for approximately 830 occupations, for the nation, states and metropolitan and nonmetropolitan areas, and also publishes national occupational estimates for specific industries. Those estimates count people in an occupation, not purchase authority, so treat them as a check that the role exists, not as a market size.
Budget and delivery fit
Budget fit asks whether the account already spends money or paid time on the problem. The current workaround is the clearest evidence: overtime, a contractor, a spreadsheet maintained by a salaried employee or a generic tool bought for something else. If nothing is spent today, the purchase has to create a new budget line, which is slower and riskier for a first customer. Ask who controls the money the workaround uses and whether that person can approve a small purchase without a formal procurement process.
Delivery fit asks whether you can serve the segment with the product you have, or nearly have, without custom work for each account. Integrations with systems the segment already runs, compliance requirements, on-site installation and support hours all count. A segment that needs a different product for each customer turns early sales into a consulting business. Write down what each candidate would require that the current product does not do.
Reachability belongs here too: you need enough accounts to learn from, and you need to be able to find them. The US Census Bureau's County Business Patterns series provides subnational economic data for establishments with paid employees by industry and employment size, including the number of establishments, employment and payroll. Its program description notes that private businesses use the data for analyzing market potential and that statistics are available by state, county and metropolitan area. A count of establishments in the right industry and size band within your reachable geography tells you whether a segment can supply a few dozen first conversations. Businesses without paid employees are not counted, and the program page listed 2023 as the latest data year when checked in October 2026, so treat the counts as a recent baseline rather than a current list.
The ICP evidence matrix
The matrix puts the candidates side by side and forces each cell to state what kind of evidence supports it. Use one column per candidate segment and grade each cell as observed, reported, assumed or unknown. The table below defines what counts.
| Criterion | Question to answer | Counts as observed | Does not count |
|---|---|---|---|
| Buying trigger | What event makes the problem urgent, and can you see it from outside? | Named accounts where the event happened recently and the buyer confirmed it mattered | A trend that will affect the whole industry eventually |
| Buyer role | Which job title approves the purchase, and how did the last related purchase happen? | A described approval path for a recent purchase | A user who says they would recommend the product |
| Budget | What does the account spend on the workaround today, and who controls it? | A workaround you saw, with its cost in money or paid time | Agreement that the problem is costly |
| Delivery fit | What would this segment need that the product does not do yet? | A short written gap list confirmed by a buyer | A belief that the product is general enough |
| Reachable accounts | Can you list real accounts and reach the buyer role? | A named account list and a working route to the buyer | A market size estimate |
| Commitment | Has anyone in the segment taken a step that costs them something? | A completed next step: data shared, an introduction to the approver or a paid pilot | Enthusiasm in an interview |
The decision rules are simple. A candidate with an unknown buyer role or an unknown budget is not chosen until those cells are filled. Among the remaining candidates, prefer the one with the most observed cells, and break ties on delivery fit, because a segment that needs custom work for each account slows every later step. Market size comes after that, as long as the segment has enough reachable accounts for the next round of tests. Write the assumptions the choice depends on next to the matrix, together with the evidence that would prove each one wrong.
A worked decision with explicit assumptions
The following is a hypothetical example with invented figures. It illustrates the method and does not describe a real company, customer or measured result.
A founder has built a mobile app that turns roof inspection photos into a damage report formatted for insurance claims. Three segments look plausible: residential roofing contractors with crews of ten to fifty people, property management companies running several hundred rental units, and independent home inspectors. Over three weeks the founder holds nine conversations with roofing contractors, five with property managers and six with home inspectors. In one of them, the founder climbs onto a hail-damaged roof with an estimator and watches the estimator circle each hail strike on the shingles in chalk, photograph it with a phone and later rebuild the photos into a report at the kitchen table that night.
| Criterion | Roofing contractors | Property managers | Home inspectors |
|---|---|---|---|
| Buying trigger | Observed: a hailstorm in the service area, visible in public storm reports and local news; six of nine contractors described a report backlog after the last storm | Reported: annual unit inspections and move-outs, not visible from outside | Weak: every inspection is routine, with no event that changes urgency |
| Buyer role | Observed: the owner approved the last software purchase at four of the contractors | Unknown: software is chosen at head office, and nobody interviewed had signed a purchase | Observed: the inspector is both user and buyer |
| Budget | Observed: estimators work evenings after storms to finish reports, paid as overtime at three contractors | Reported: staff time on inspections, cost not tracked | Observed but small: inspectors already pay for report software they are satisfied with |
| Delivery fit | Fits the current app; one contractor asked for a branded report cover | Needs integration with the property management system | Fits, but competes with established report tools |
| Reachable accounts | A county-level establishment count suggests enough contractors in the founder's region for the next round | Few accounts in reach, each with a long approval cycle | Many accounts at a low price |
| Commitment | Two contractors shared the photo folder from a full storm; one agreed to a paid pilot for the next storm | None beyond interviews | One inspector agreed to try a free version |
Property management is excluded by the decision rule because the buyer role is unknown. Home inspectors have an observed buyer and a clean fit, but the trigger is weak and the budget already goes to an incumbent they are satisfied with. Roofing contractors have the most observed cells and a completed paid commitment, so the founder chooses them as the first ICP: residential roofing contractors with crews of ten to fifty in the founder's region, triggered by a hailstorm in their service area, with the owner as the approving buyer.
The choice rests on three written assumptions, each with a test that could prove it wrong:
- Storms are frequent enough. The founder assumes the region gets enough hail events during the test period to produce triggers. If no qualifying storm occurs within the first two months, the trigger is too rare for a steady pipeline.
- Owners approve without procurement. If more than two of the first five pilots stall in an approval step the owner did not mention, the buyer role needs another look.
- No custom work. If more than one pilot needs a change beyond branding, delivery fit was overstated.
After the next storm, the founder contacts ten contractors in the affected area within a week and records how many accept a call, how many share photos and how many pay. Those counts, not the size of the roofing market, decide whether to keep the ICP, narrow it further or move to the next candidate.
Common mistakes and limits
- Choosing by market size. A large total market says nothing about whether any account will act this quarter.
- Grading reported evidence as observed. What a buyer says happened is useful, but a workaround you saw or a step the buyer completed is stronger.
- Describing a department instead of a role. A department does not sign contracts; a person with a job title does.
- Keeping several ICPs at once. Splitting outreach across segments makes each test too small to read.
- Treating the choice as permanent. The first ICP is a hypothesis for the next round of selling, not an identity.
The matrix has limits. It suits products sold to identifiable organizations, where a purchase has a buyer and an approval path. Consumer products and self-serve tools rely more on observed purchase and usage data than on interviews. Public statistics can tell you whether enough accounts exist, not whether any of them will buy. The matrix and its decision rules are an editorial framework, not a measured result.
Related guides and next step
- Placing the ICP in a wider plan: How to Build a Startup Go-to-Market Strategy by Stage
- Collecting the interview evidence: How to Run a Customer Interview Without Pitching
- Choosing questions that reveal intent: Customer Discovery Questions That Reveal Buying Intent
- Testing a segment in a new country: How to Test a US Beachhead Before Opening an Office
Daniel's go-to-market mentoring page describes the first step of his process as defining a narrow ICP, a buying trigger and a decision-maker, in his words: 'One segment, one moment that makes them buy, and the exact job title that signs.' It describes selecting accounts on signals such as hiring, funding, regulatory changes, leadership moves or visible evidence of the manual process being replaced. That account is self-reported on his site.
Bring your evidence matrix and interview notes to a free 30-minute strategy session to review your ideal customer profile decision with your actual evidence and constraints.
Frequently asked questions
How do you choose your first ideal customer profile?
List two or three candidate segments and grade each on the same evidence: an observable buying trigger, a named buyer role with spending authority, money or paid time already spent on the problem, delivery without custom work, reachable accounts and any completed commitment. Exclude candidates with an unknown buyer or budget, then pick the one with the most observed evidence.
What evidence should a founder collect for an ideal customer profile?
For each candidate segment: accounts where the trigger happened recently, the approval path of a recent related purchase, the cost of the current workaround, a written list of product gaps, a named account list and any step a buyer completed, such as sharing data, introducing the approver or paying for a pilot. Record whether each item was observed, reported or assumed.
What are common mistakes when choosing an ideal customer profile?
Choosing the segment with the largest market, treating what buyers said as if it were observed, naming a department instead of a job title, pursuing several segments at once and treating the first choice as permanent rather than as a hypothesis for the next round of selling.
How narrow should a first ideal customer profile be?
Narrow enough that you can list real accounts by name, state the job title that approves the purchase and describe the event that makes those accounts act. If you cannot list accounts, the description is still a market, not an ICP.