In venture-backed tech companies, there is a phase where the conversation shifts from “how do we serve this customer better” to “how do we find more customers, or different ones.”
The second question feels like ambition. It often signals something else.
In the phase after initial traction, when early PMF is established but the growth model is still being stress-tested, many companies begin expanding their ICP. The sales team starts closing deals with profiles that don’t quite match the original. Marketing adjusts messaging to cast a wider net. Product begins fielding requests from new segments and adding features to serve them. Each individual decision has internal logic.
One founder described the moment of recognition clearly: “You go from knowing all your customers to knowing none of them.”
What This Usually Looks Like
The pattern often starts in the sales org. A deal closes with a company that’s slightly outside the original profile . . . different size, different use case, different vertical. The team celebrates the new logo. Then another edge deal closes. Then another.
Marketing watches the expansion and adjusts. Messaging broadens to avoid excluding potential buyers who don’t match the original profile. The ICP document gets quietly revised, or quietly stops being referenced altogether. Product inherits feature requests from the new segments and begins building for them.
The leadership team typically frames this as market development: learning where else the product creates value, expanding the addressable market, not narrowing prematurely. None of these explanations are wrong on their own. The problem is what they obscure.
What’s Actually Happening
When original fit is strong, companies don’t typically expand the ICP deliberately . . . they narrow toward it. They say no to outliers because the core converts predictably, with lower CAC, faster cycles, and stronger retention. There is enough signal in the original segment to trust.
ICP expansion at this stage is rarely a strategic move. It is a pressure release.
When the original segment starts producing less reliably - not catastrophically, but inconsistently - the instinct is to cast wider. More profiles, more use cases, more verticals. The expansion gets framed as opportunity. It is more often a response to weakening fit in the original market.
This is one of the most common forms the pattern takes that I described in last week’s post When Motion Replaces Clarity: simultaneous change can look like innovation, but it also signals the growth model has begun to drift, where leadership teams add motion when clarity drops. ICP expansion is that motion with a market-development narrative attached to it.
The Leadership Shift
The question that matters isn’t “which new segment should we enter?” It is “what is actually happening in the segment we already understand?”
When the original ICP was working, that question had a clean answer. The profile was clear, the motion was repeatable, and the team could describe the ideal customer without significant disagreement. When that question stops having a clean answer, expansion gets proposed as the alternative.
But expansion doesn’t restore fit. It dilutes the system. Every edge case that closes teaches the sales team a slightly different motion. Marketing tests a slightly different message. Product inherits a slightly different set of requirements. Over several quarters, the growth system that was once coherent now serves multiple overlapping profiles — none of them deeply enough to produce consistently repeatable outcomes.
Why This Matters
The cost of ICP drift compounds slowly, which is part of why it’s difficult to catch early. In the short term, expansion looks like progress. Pipeline grows. New logos appear. The team has new conversations and new energy.
Over time, conversion rates begin to soften . . . not uniformly, but inconsistently, which makes diagnosis harder. Sales cycles lengthen. The ideal customer profile becomes harder to defend in hiring conversations. New reps take longer to ramp because there is no single profile to anchor around.
By the time the cost shows up clearly in the numbers, the organization has usually been absorbing the friction for several quarters. The expansion that felt like acceleration was, in many cases, the growth model losing coherence one edge case at a time.
The Diagnostic Signal
One question often clarifies the picture:
If you asked your sales team to describe your best-fit customer today, would they give the same answer they would have given twelve months ago?
If yes and the profile is still converting cleanly, the model is intact. If the answers diverge, or if different functions give meaningfully different answers, ICP drift has likely already begun. That divergence is usually downstream of something the original segment stopped doing reliably, not upstream of the growth the expansion is meant to create.
Broadening the ICP before diagnosing the original fit is not market development. It is, in most cases, growth model drift moving forward under a different name.
If your team is spending more time debating who the customer is than serving them, you may already be in a growth recalibration phase. This is where leadership decisions begin to change.