Your first ICP: how to pick who to sell to before you have data
Pick your first ICP before data by using pain, triggers, buyer access, and live sales evidence instead of broad market guesses.
Short answer. Pick your first ICP by choosing a painful use case and a reachable buyer with a visible trigger. Early data is thin, so the job is to make a sharp selling thesis. I would start with companies that recently exposed the problem in hiring, regulation, funding news, tool sprawl, or support tickets. Then I would test whether the same words make strangers reply.
In short:
- Your first ICP is a falsifiable sales thesis: buyer, pain, trigger, and reachability.
- Firmographics matter after pain; start with the moment that makes a buyer search for a fix.
- A narrow wedge beats a broad market because outbound needs a reason to land in the right inbox.
- Change the ICP when reply quality and budget language point to a tighter buyer.
What is a first ICP?
Start narrower. A first ICP is the smallest believable set of accounts where the same painful event and buying path repeat enough to guide outbound before a dashboard exists. It is a bet you can write on a whiteboard. My test is simple: can I name the company and the person carrying the pain, then point to the trigger that makes this quarter matter?
Steve Blank's customer development frame is useful because a startup is built to search before it scales. That matters for an ideal customer profile for a startup. The early profile should create conversations, not decorate a pitch deck. If company size or geography cannot change the next email I send, I would keep it out.
- Pain named in buyer words
- Trigger visible from outside
- Buyer reachable by founder
- Deal path plausible this quarter
Where do you find clues before data?
Look for public smoke. Before a CRM has patterns, the best clues sit in job posts, incident write-ups, new compliance pressure, marketplace reviews, help forums, and competitor complaints before any salesperson arrives. Those nouns tell you who is already spending attention. I trust observed mess more than a survey checkbox because founders usually overstate what a polite prospect will do.
The job-to-be-done lens from Clayton Christensen helps because it asks what progress the buyer is trying to make in a real situation. For a B2B startup, I would translate that into a sentence a CFO, RevOps lead, plant manager, or security owner might actually say. The ICP begins where that sentence sounds expensive.
- Hiring spike around the problem
- New tool listed in stack
- Support backlog in reviews
- Regulatory notice on the calendar
- Manual spreadsheet still visible
Choose a wedge before a market
Pick the wedge. A market label like healthcare, fintech, logistics, or SaaS sounds mature, yet it rarely tells me why a buyer opens an email from a stranger today. A wedge has a named workflow with pressure around it. The ICP gets useful when I can say which accounts are bleeding time inside that workflow and which title owns the mess.
I would rather define the ICP as mid-market freight brokers drowning in carrier onboarding than companies in logistics. Specificity changes the sales motion. It shapes the subject line, proof point, demo path, pricing anchor, and event list. Broad language gives the founder false comfort because almost every account can appear relevant from far away.
Who feels the invoice pain?
Follow the invoice. The first ICP should sit near the person whose budget, bonus, board slide, or weekend gets damaged by the problem when the current workaround breaks during live work. Users matter, but unpaid enthusiasm can waste a founder's month. A clean ICP names the economic buyer and the internal champion, then accepts that they may speak in different nouns.
In founder-led sales, the early mistake is treating every friendly user as proof of an account profile. I care about who can move money. The best clue is the phrase a buyer uses when the problem reaches a vendor meeting, finance review, security review, or renewal call. That phrase belongs in the ICP.
- Economic buyer name
- Champion's daily wound
- Approval path
- Budget source
How do you test the ICP without a dashboard?
Use live selling. A first ICP becomes real only when strangers in that segment react with recognition or a concrete objection instead of offering soft encouragement to end the conversation. I do not need volume at the start; I need signal density through replies that mention the same pain and calls that reach the same fork. Related objections matter more than random politeness.
The test can be a cold email, a founder note after an event, a LinkedIn message, or a short problem call. Keep the asset plain. If the buyer needs a polished deck to understand the pain, the ICP may still be too vague. A sharp profile produces sharper objections, which is useful even when the answer is no.
- Reply names the pain
- Buyer corrects the wording
- Call reveals budget owner
- Objection repeats across accounts
- Referral points sideways
When should the ICP change?
Move when evidence moves. I would change the ICP when a different segment shows sharper pain or a faster path to budget while the current segment keeps giving vague praise and slow calendars. The danger is emotional switching after a bad week. A pivot should come from repeated sales evidence, with boredom kept out of the room.
Hold the line long enough for the market to answer. Too many founders rewrite the ICP after every awkward call, which turns learning into weather. I would separate message problems from market problems. If the same buyer understands the pain but refuses urgency, the trigger may be wrong while the segment still deserves another pass.
Common questions
How do I define my ideal customer profile for my startup?
Define it as a narrow selling thesis around a painful workflow and a reachable buyer. A useful startup ICP names the account type, the person with budget pressure, the trigger visible from outside, and the reason the problem is urgent now. Leave out details that do not change your email or discovery call.
How do I create an ideal customer profile for my B2B startup?
Create it by starting from the painful moment, then adding firmographics only when they explain why that moment appears. For a B2B startup, I would write a short ICP sentence with account type, trigger, buyer title, current workaround, and likely budget source. The profile earns trust when outbound replies repeat the same pain language.
How do I identify my ideal customer profile (ICP)?
Identify it by watching where pain becomes visible without your product. Job posts, public complaints, migration projects, policy deadlines, and renewal pressure can reveal accounts already feeling the issue. Then compare conversations. The ICP is the segment where buyers understand the problem quickly and can describe what happens if nothing changes.
What should be in a first ICP?
A first ICP should include account shape, painful workflow, buyer title, trigger, current workaround, and disqualifiers. I would also include the words the buyer uses for the problem. The disqualifiers matter because they protect founder time. If an account looks impressive but lacks urgency or reachable ownership, it belongs outside the early profile.
How narrow should my first ICP be?
Narrow enough that the same outbound message feels written for the buyer's week. If accounts need separate pain language and separate proof, they are probably different ICPs for now. A narrow ICP should still have reachable accounts, visible triggers, a budget owner, and timely pressure. Tight beats vague because founder-led sales needs precision.
When should I change my ICP?
Change it when repeated sales evidence points elsewhere. That means sharper replies, clearer budget ownership, shorter education, or more urgent trigger events in another segment. Do not switch because a bad call stung. Message quality and founder delivery can fail even when the underlying ICP is sound, so separate wording problems from market problems.