Who Are You Actually Selling To?

ICP stands for Ideal Customer Profile. You’ve probably heard the term. You’ve probably even written one down in a pitch deck. “We sell to mid-market B2B SaaS companies.” That is not an ICP. That is a census category. It tells you nothing useful about who to call on Monday morning.

A real ICP is specific enough that you could walk into a room of 100 people and within five minutes identify the 3 who would buy your product this week. Not this quarter. This week.

A real ICP answers five questions. Who is the person (their role, not their company)? What stage is their company at? What is the specific trigger event that makes them need your product right now? What pain are they feeling daily? And what are they currently doing about it that isn’t working?

When all five of those line up, you don’t have a prospect. You have a customer who just hasn’t signed up yet. The conversation is easy. The demo is obvious. The close is fast. Not because you’re a great salesperson, but because you found someone whose hair is on fire and you’re holding a bucket of water.

When even one of those five is off, the conversation is a slog. You’re pushing. They’re hesitant. You leave the call wondering what went wrong. Nothing went wrong. You were just talking to the wrong person.

This is the single most useful filter for early stage sales. Forget market size. Forget TAM. Forget how many companies match your firmographic criteria on LinkedIn Sales Navigator. The only question that matters at your stage is: is this person’s problem so painful that they would pay to fix it today?

Not “they agree it’s a problem.” Not “they said it would be nice to solve.” Not “they put it on their roadmap for Q3.” Today. Right now. If you called them and said “I can make this problem disappear, it costs $99 a month,” would they give you a credit card on the phone?

That is the hair on fire test. And most of the people you’re talking to will fail it.

This is fine. This is actually the point. The goal of early ICP work is not to find as many prospects as possible. It’s to find the smallest, most specific group of people who have an urgent problem and are actively looking for a solution. You’re not casting a wide net. You’re looking for the 30 people on earth who need your product more than anyone else, and you’re calling every single one of them.

I know “smallest possible group” sounds counterintuitive. Your investor wants to hear about a billion dollar market. Your pitch deck has a big TAM circle. But the TAM slide is for fundraising. The ICP is for selling. And in selling, narrow beats wide every single time at your stage.

Here is why. When you sell to a narrow ICP, your messaging is razor sharp because everyone has the same problem. Your demo is fast because you know exactly what to show. Your close rate is high because you’re only talking to people who need what you have. Your product gets better faster because all your feedback comes from similar users with similar needs. And your customers talk to each other, which means word of mouth kicks in naturally.

When you sell to a broad ICP, everything is mediocre. Your messaging is vague because you’re trying to speak to ten different personas. Your demo is long because you don’t know what they care about. Your close rate is low because half the people you talk to have a mild inconvenience, not a burning problem. Your product gets pulled in ten directions. And your customers have nothing in common so nobody refers you to anyone.

Narrow down until it feels uncomfortable. Then narrow down one more time. That’s your ICP.

Force yourself to write it in one sentence. If you can’t, you don’t know it well enough yet.

The format is: [Role] at [company type/stage] who [trigger event] and is currently [broken workaround].

Example: “A B2B founder at a seed to Series A startup, team of 2 to 10 people, who is doing 3 or more external sales calls per day and has abandoned their CRM because nobody on the team was updating it.”

That sentence tells you exactly who to look for. You know their role (founder). You know their company stage (seed to Series A). You know the team size (small). You know the trigger (high call volume). And you know the current state (CRM is dead, data is a mess).

With that sentence, you can scroll through a list of YC companies and within seconds know who to reach out to and who to skip. You can get on a call and within the first two minutes confirm whether this person matches. You can write outbound emails that feel personal because they describe a situation the person is actually living.

Without that sentence, you’re guessing. And guessing at your stage means wasting the only resource you don’t have enough of: time.

Here is a test. Read your ICP sentence out loud. If it could describe more than 50,000 companies, it’s too broad. If it could describe fewer than 200, it might be too narrow (but probably isn’t, honestly). The sweet spot for your first ICP is somewhere between 500 and 5,000 companies. Small enough to be specific. Large enough that you won’t run out of people to call.

This is the part most founders skip, and it’s just as important as the ICP itself. Your anti-ICP is the explicit list of people you will not sell to. Not because they’re bad people, but because selling to them will waste your time, pollute your feedback, and slow everything down.

Your anti-ICP should make you uncomfortable. It should include people who would probably take a meeting with you. People who might even say yes. But who will ultimately churn, ask for features you shouldn’t build, or take 6 months to close when your average cycle should be 2 weeks.

Some common anti-ICP patterns for early stage B2B startups:

Companies with more than 50 people on their sales team. They have different problems than your ICP. They need enterprise features, SSO, admin controls, custom reporting. They’ll ask for all of it during the sales process and you’ll spend three months building things that don’t help your core customer.

Companies whose primary motion is mass outbound email. If they’re sending 10,000 cold emails a month, they need a different kind of tool. They need deliverability management, email warming, sequence builders. That’s a different product for a different buyer.

Companies that are “just exploring.” If they don’t have an active pain, if they’re “curious about what’s out there,” if they want to “keep you in mind for later,” they are not your customer right now. Be polite, be helpful, and move on.

Companies where the person you’re talking to can’t make a buying decision. If they need to “run it up the flagpole” and you can’t get on a call with the actual decision maker, you’re going to spend weeks waiting for an answer that’s probably no.

Write your anti-ICP down. Put it somewhere you’ll see it before every sales call. When you catch yourself on a call with someone who matches the anti-ICP, end the call politely and move on. Do not convince yourself that “this one might be different.” It’s not different.

I cannot overstate how much time this saves. Every minute you spend with the wrong prospect is a minute you didn’t spend with the right one. At your stage you might only have capacity for 5 calls a day. If 3 of those calls are with anti-ICP prospects, you’re operating at 40% efficiency. Fix the filter and everything downstream gets better.

Of the five ICP components, the trigger event is the most underrated and the most powerful. A trigger event is the specific thing that happened recently that made this person’s problem go from “annoying” to “urgent.”

Without a trigger event, you’re calling someone and saying “hey, you probably have this problem.” They say “yeah, kind of” and the conversation goes nowhere because it’s not a priority today.

With a trigger event, you’re calling someone at the exact moment they need you. The conversation has energy. They’re leaning in. They want to solve this now, not next quarter.

Trigger events for a CRM product might look like: they just hired their second or third salesperson and realized nobody is using the same system. They just lost a deal because a follow-up fell through the cracks. They just did a board meeting and their investor asked about pipeline and they had to admit they don’t actually know. They just tried HubSpot for the third time and rage-quit after spending a weekend on setup. They’re a YC company and Demo Day is in 8 weeks and they need to show traction.

Each of these trigger events creates a window. The window might be open for two weeks or two months. But it will close. The pain will get normalized. They’ll find a hacky workaround. They’ll convince themselves it’s fine. Your job is to find them while the window is open.

How do you find trigger events? Some of them are visible. Job postings for sales roles mean the team is growing. A new funding round means they have money to spend and pressure to grow. A new CRM-related complaint on Twitter means someone is frustrated right now. LinkedIn posts about “fixing our sales process” mean someone is in the window.

Some trigger events you discover on the call itself. That’s fine. The first question of every call should be something like “what made you take this meeting?” or “what’s going on right now that made this a priority?” If they can’t answer that question clearly, they probably don’t have a trigger event, and the deal is going to stall.

The ICP you define today will be wrong. Not completely wrong, but wrong enough that you’ll need to revise it. Probably multiple times in the first year.

This is not a failure. This is the system working correctly.

Your first ICP is a hypothesis. You test it by having conversations. Some conversations confirm the hypothesis. Some contradict it. After 50 conversations, you revise. After 100, you revise again. The ICP gets sharper every time.

The founders who struggle are the ones who treat their first ICP as permanent. They defined it in their pitch deck, their investors nodded, and now they feel locked in. “We sell to mid-market fintech companies” becomes an identity, and they keep grinding on that ICP even when the data is screaming at them to pivot.

Stay loose. Pay attention to which deals close fast and which deals drag. The deals that close in under two weeks are telling you who your real ICP is. The deals that take three months are telling you who your anti-ICP is, even if they eventually close. A three month deal at your stage is not a win. It’s a warning.

Look at your best five customers. Not your biggest five. Your best five. The ones who activated fast, who use the product every day, who would be upset if you disappeared. What do they have in common? That commonality is your real ICP. It might be completely different from what you wrote in your pitch deck. That’s fine. The pitch deck was a guess. The data is real.

Get a spreadsheet. Just a simple one, nothing fancy. Make columns for: company name, contact name, role, company stage, team size, trigger event, current workaround, and outcome (did they buy, and how fast).

Fill this in for every sales conversation you have. After 30 rows, sort by outcome. Look at the people who bought fast. Really look at them. What do they have in common that the slow deals and lost deals don’t?

That pattern is your ICP. Write it down in one sentence. Put it at the top of the spreadsheet. Before every new call, ask yourself: does this person match the sentence? If yes, take the call. If no, politely decline or deprioritize.

Revisit the sentence every month. It will evolve. Let it.

Most founders resist this exercise because it feels too simple. They want a sophisticated framework with weighted scoring and persona matrices. They want to feel like they’re doing serious strategy work. The spreadsheet feels beneath them.

The spreadsheet is the strategy. Everything else is theater.