18,000 Companies Past Their Exit Window: What I Keep Seeing in the GTM Data
McKinsey's 2026 Global Private Markets Report puts a number on something operating partners already feel: 18,000 portfolio companies are past their traditional exit window, with average holding periods at 6.7 years — the longest since 2005. I'm not going to pretend I have a PE fund's worth of portfolio data to validate that independently. But I've spent enough time inside the GTM layer of growth-stage companies to recognize the pattern McKinsey is describing, and it's not primarily a rate environment problem.
The companies that can't exit aren't failing because buyers are cautious. They're failing diligence on metrics that should have been managed two years earlier.
The Diligence Killers Are GTM Problems
The Bain/StepStone 2026 GP Survey names poor earnings quality and customer churn as the top two reasons PE deals fall apart in diligence. I want to be precise about what that means operationally: earnings quality degrades when you've been selling to the wrong customers. Churn accelerates when your value proposition drifted from the people who actually needed the product. Both of those are ICP problems. Both of them show up in the data room long after the window to fix them has closed.
I've watched this happen at the company level — not the portfolio level, I'll be honest about that limit. But the mechanism is the same whether you're looking at one company or thirty: when customer definition is loose, retention is unpredictable, and unpredictable retention is what kills a multiple.
The NRR Gap Is Real and It's Quantifiable
The link between net revenue retention and exit multiple has been documented: 113% NRR corresponds to roughly a 24x revenue multiple; 98% NRR corresponds to roughly 5x. I can't independently verify those exact figures — they come from public SaaS benchmarking data — but the directional reality matches what I've seen. The spread between a company with strong expansion motion and one with flat retention isn't marginal. It's the difference between a fundable narrative and a hold that grinds into year seven.
What I can say from direct experience: most companies don't know their real NRR. They know their reported churn. Those are not the same number, and the gap between them is usually a GTM problem — customers who were never properly onboarded to the use case that would make them expand.
The Comparability Problem Is Real, But I'll Be Specific About What I Know
I've heard operating partners describe the portfolio comparability problem firsthand: every company reports revenue health in its own language, with its own definitions, on its own cadence. That fragmentation makes it genuinely hard to identify which companies have exit-grade revenue quality and which ones need intervention before they enter a process.
I'm not going to claim I've built a portfolio-wide diagnostic system — I haven't. What I've built is a way to run that diagnostic at the company level, consistently, so the outputs are comparable if you choose to apply it across holdings. Whether that scales to a fund-level operating capability depends on how the firm wants to use it. I'm honest about that boundary.
What Actually Moves the Number
If I were advising an operating partner on where to focus, based on what I've seen work at the company level:
Get the ICP documented before it lives only in one person's head. The VP of Sales who defined the customer profile may not survive the transition. If that knowledge isn't written down and legible to an incoming team or acquirer, you've got a diligence risk that doesn't show up until it's too late.
Watch leading churn signals, not lagging reported churn. By the time churn shows up in the number, the cause is six months old. The signals that predict it — support ticket patterns, usage drop-off, stakeholder turnover at the account — are visible earlier if you're instrumented for them.
Build the board narrative around revenue quality, not just growth rate. Buyers are pricing durability. NRR, customer concentration, expansion motion — these belong in the board deck throughout the holding period, not assembled for the first time in the data room.
I'll be direct about what I don't know: I haven't run a portfolio-wide GTM transformation. I've worked at the company level, and I've talked to enough operating partners to understand the structural problem they're describing. If you're an operating partner thinking about this, I'd rather have an honest conversation about where Andru's diagnostic actually fits than oversell a capability I haven't proven at your scale.
The 18,000 companies stuck past their exit window aren't all fixable. But the ones that are fixable have a GTM problem at the root — and that's a problem with a known mechanism.