How Do You Pick a Second ICP Segment After Your First Ten Customers?

How do you pick a second ICP segment after your first ten customers?

Look at your first ten customers and find the group that bought fastest, needed the least custom work, and is still using the product. Then pick the nearest neighbour to that group, a segment that shares the same pain but differs on one variable such as company size, industry, or buyer role.

Most founders I talk to treat the second segment as a bet on a bigger market. That is usually the wrong frame. The second segment is a test of whether your first wins were a pattern or a coincidence. If you change too many variables at once, you will not know which one broke when deals stall.

I work with B2B SaaS teams on go-to-market strategy and the systems that run it, and this decision comes up right after the founder-led sales phase. It deserves more care than it usually gets, because the second segment shapes your messaging, your outbound lists, and your hiring plan for the next year.

Why is the second ICP segment harder than the first?

The first segment usually comes from the founder's network and instinct, so it works partly through trust and personal context. The second segment has to work without that advantage. It tests whether your positioning, proof, and sales motion can win buyers who have never heard of you.

Your first ten customers often share something you did not design. Maybe they came from one Slack community, one former employer, or one investor's portfolio. That shared context makes deals easier, and it hides weaknesses in your pitch. When you step outside it, cold buyers ask harder questions and expect proof you may not have yet.

This is why I push teams to be honest about where each early customer came from before choosing what comes next. A warm introduction is not the same as a repeatable channel. The second segment should be chosen with that difference in mind.

What should you look for in your first ten customers?

Look for three signals in each account: how long the deal took from first call to signature, how much custom work onboarding required, and whether the customer is still active and growing. The accounts that score well on all three show you the real shape of your ideal customer profile.

I like to put the ten accounts in a simple Airtable base or a HubSpot view with a handful of columns. Industry, employee count, buyer title, source of the deal, days to close, onboarding effort, and current health. Nothing fancy. The goal is to see the pattern with your own eyes rather than argue about it in a meeting.

Then add one more column that most teams skip: the trigger. What was happening at the customer when they bought? A new hire, a failed tool, a funding round, a compliance deadline. The trigger often tells you more about the next segment than firmographics do, because a trigger travels across industries in a way that company size does not.

What is the nearest neighbour approach to ICP expansion?

The nearest neighbour approach means changing exactly one variable from your winning segment while keeping everything else the same. If your best customers are fifty-person fintech companies buying through the head of operations, try fifty-person insurtech companies with the same buyer, not enterprise banks with a different one.

The power of this approach is diagnostic clarity. When you change one thing, a result tells you something. If insurtech converts like fintech, your positioning travels across that industry line. If it stalls, you know the industry context mattered, and you can dig into why with a handful of discovery calls.

Compare that to jumping straight to a new size band, a new industry, and a new buyer role together. If that test fails, you learn almost nothing. You just lose a quarter. I wrote about the related problem of choosing between two segments that both buy in how to pick an ICP when two segments both buy, and the same discipline applies here.

Which variable should you change first?

Change the variable that keeps your proof most relevant. For most early B2B SaaS teams that means staying with the same buyer role and similar company size, and moving to an adjacent industry. Buyers trust case studies from companies that look like them, so the closer the match, the more your existing proof still works.

Here is how I rank the usual options. Adjacent industry with the same buyer is the safest, because your demo, your pain language, and most of your case studies still land. Same industry with a larger company size is next, but it often brings procurement, security reviews, and longer cycles you may not be ready for.

A new buyer role is the riskiest change. A product sold to marketing operations and the same product sold to the chief financial officer are almost two different businesses. The pain is described differently, the budget lives somewhere else, and the objections change. I would only make that move once the first segment is clearly repeatable.

How do you test a second segment without derailing the first?

Run the test as a time-boxed experiment with its own list, its own messaging, and a fixed share of selling time. Protect the first segment's pipeline, set a clear success threshold before you start, and review results on a set date rather than letting the experiment drift on indefinitely.

In practice, that means building a focused list of accounts in the new segment using a tool like Clay or Apollo, enriching it with the same fields you used for your first ten customers, and writing a sequence that speaks to that segment's version of the pain. Keep it small enough that a founder or one rep can work it properly.

Decide up front what counts as a pass. It might be a certain number of discovery calls that confirm the pain, or a first paid pilot within the window. Write it down. Teams that skip this step end up keeping a weak segment alive because nobody wants to call it, which is a pattern I cover in how to retire an ICP segment that stopped converting.

What mistakes do founders make when choosing a second segment?

The most common mistakes are chasing the largest market instead of the closest one, letting one big inbound lead define the segment, and changing positioning for everyone before the new segment has proven itself. Each one trades a clear signal for a hopeful story.

The single big lead is the sneakiest. An enterprise logo fills out your demo form, the founder gets excited, and suddenly the whole roadmap bends toward that segment. One lead is an anecdote. Before it shapes your GTM plan, you need to see whether five or ten similar companies feel the same pain.

The other mistake I see often is rewriting the homepage for the new segment on day one. Your website should keep serving the customers who already convert. Build a dedicated landing page or a segment-specific sequence for the test, and only change your core positioning once the new segment has earned it.

How should your GTM systems change for a second segment?

Add a segment field to every account and contact in your CRM, route and report by that field, and keep separate sequences and lead scoring rules per segment. Without that separation, you cannot tell whether the new segment is working, because its numbers blend into the first one.

This is where GTM engineering earns its keep. In HubSpot, a simple dropdown property on the company record, filled by enrichment or by the rep, lets you split pipeline, win rate, and cycle length by segment. Lead scoring should also be segment aware, because the job titles and company sizes that signal intent in one segment may mean nothing in the other.

Reporting is the part teams forget. Set up one dashboard that shows both segments side by side: meetings booked, opportunities created, win rate, and average days to close. Review it on the same cadence as your pipeline meeting. If you started from your closed-lost data, the approach in how to define an ICP from closed-lost deals pairs well with this view.

What should you do next?

Build a one-page view of your first ten customers with deal speed, onboarding effort, health, and trigger. Pick the nearest neighbour segment that changes one variable. Then run a time-boxed test with its own list, sequence, and success threshold, and tag everything by segment in your CRM.

The second segment is not about getting bigger faster. It is about learning which parts of your early success will travel. Get that right and the third and fourth segments become much easier, because you will know which variables matter and which do not.

If you want a second pair of eyes on your first ten customers, or help setting up the segment tracking in HubSpot, reach out through pravinkumar.co. Let's chat about where your next segment should come from.