Roughly seven out of ten CEOs of private-equity-backed companies do not survive the hold. Most of the replacements cluster between months 18 and 24 — late enough that the fund has burned a third of its runway, early enough that almost nothing irreversible has actually gone wrong with the business. I have watched this movie from every seat: as a CEO reporting to sponsors, as a chairman hiring and replacing CEOs, and as the operator brought in after the change. The pattern is so consistent it should embarrass our industry, because the failure is almost never the person. It is a set of misalignments that were visible — and fixable — in the first hundred days.
The number nobody puts in the deck
Start with what the 70% actually costs, because sponsors systematically underprice it. The direct costs — search fees, severance, sign-on packages — are rounding errors. The real bill: six to nine months of drift while confidence erodes and the board manages instead of governs; three to six months of search; six months of the new CEO learning what the old one already knew. Call it a lost year to eighteen months on a six-year hold — 20 to 25 percent of the entire compounding window, spent going sideways. On a fund model that needs EBITDA growth to carry the return, a lost year is frequently the difference between the 3x you underwrote and the 2.1x you report. And that is the cost when the replacement works. When it doesn’t — and second CEOs fail at alarming rates for exactly the same reasons as first ones — the deal is usually written down for good.
Misalignment one: the CEO doesn’t know the number
Ask the CEO of a portfolio company what MOIC the fund underwrote and watch the pause. In my experience the majority cannot answer — not because they are incurious, but because nobody ever translated the fund’s math into an operating target and put it in writing. The deal team knows the number. The IC memo knows the number. The CEO got a budget. A budget is an annual, negotiated, incremental document; a fund model is a multi-year compounding commitment. The two diverge from day one, invisibly, because they are denominated in different units. By month 18 the budget has been met twice while the bridge to the exit number has quietly fallen a year behind — and the board experiences that gap as the CEO’s failure, when it is actually the board’s translation error. The fix costs one meeting: the board’s number, stated, agreed, documented, in the first hundred days. I call it the Board Lock, and I will not run a company without it.
Misalignment two: the plan was never honest about the team
Every deal closes with warm words about the management team, and every sponsor privately knows one or two seats are stretched. The mistake is not the optimism; it is that nobody writes the truth down while it is still cheap to act on. The honest audit — which seats can carry the plan, which need support, which need to change — is skipped because the first quarter is busy and the conversation is awkward. So the stretched seat stays stretched, the CEO compensates by absorbing that function personally, their real job stops getting done, and by month 15 the board sees a CEO who is ‘in the weeds’ and ‘not strategic.’ They are in the weeds because they are doing two jobs — one of them the job the audit would have flagged. When I take a chairman’s seat, the capability audit against the plan happens in the first sixty days, in writing, with a support-or-change decision on every gap. It is an act of respect, not brutality: strong operators hear they are trusted, stretched ones get help before they fail publicly, and the CEO stops paying interest on a debt the whole board agreed to ignore.
Misalignment three: the board sees results, not the plan
The third failure is visibility. Most portfolio companies report results monthly — revenue, EBITDA, cash — against budget. Results are lagging indicators of a plan the board has usually never seen in operational detail. So every variance becomes a surprise, every surprise becomes a narrative, and after enough narratives the board stops believing the storyteller. That is what a confidence crisis actually is: not a performance judgment but an information failure that compounds into one. The alternative is bridge reporting: every month, actual EBITDA against the value-creation bridge, lever by lever, in dollars, with the owner of each lever reporting their own line and the correction already attached to any miss. When a board reads that report for a year, something changes in the relationship — misses stop being character evidence and become engineering problems. I have never seen a CEO fired over a miss the board understood three weeks before the meeting. I have watched several fired over misses they explained for the first time in the meeting.
The 18-month trap: why it always breaks there
The clustering at months 18–24 is not random. Month 6 is still honeymoon; the plan is ‘ramping.’ Month 12 delivers the first annual results, usually near budget, and everyone relaxes slightly. But somewhere in the second year the fund’s own clock intrudes: the sponsor starts previewing the exit math, converts the current run-rate into an implied outcome, and discovers the compounding that budgets hid. The gap that was 4% a year for two years is suddenly 30% of the equity case. Now the sponsor faces a brutal decision tree: push the CEO (they’ve ‘already been told’), change the plan (admits the underwriting), or change the CEO — the only option that feels like action. Month 18 to 24 is exactly when a search can still ‘save the deal’ in the IC’s mind. One more year of drift and the same board will conclude it is too late to change anything. The trap closes on schedule because the inputs were set at the beginning.
What survival looks like — the Three Locks
The CEOs who last the hold are not better operators than the ones who don’t, in my observation. They are better aligned, earlier. Concretely, three locks, closed in the first hundred days. The Board Lock: the exit number, agreed in writing at board level, revisited quarterly — so the CEO manages to the fund’s math, not a budget. The Capability Lock: the honest team audit, delivered without euphemism, with support and changes actioned in quarter one — so the CEO is never secretly doing two jobs. The Team Lock: compensation wired to the bridge, so every lever owner is paid on the lever they own and the plan has twenty stakeholders instead of one defendant. Add bridge reporting so the board watches the plan and not just the results, and the month-18 conversation becomes what it should have been all along: a review of a shared engineering problem.
What sponsors should do instead of firing
None of this argues that no CEO should ever be replaced — some seats are simply wrong, and the audit exists to say so early. It argues that the replacement decision, when it comes, should be a conclusion from a working system rather than a substitute for one. Before any board reaches for the search firm, three questions: Does this CEO know the number — did we ever actually lock it? Was the plan honest about the team, or has this person been compensating for a gap we all knew about? And has the board been watching the plan, or just the results? If any answer is no, the fund is about to spend a year of hold and a million dollars replacing a symptom. Fix the system first. If the CEO fails inside a locked, honest, visible system — then the seat is truly the problem, the decision takes one meeting, and the successor inherits a plan instead of a mess. That last part is the quiet payoff: even the replacements go better inside the system. The new CEO who walks into a locked number, an honest map of the team, and a year of bridge reporting is productive in month one, not month seven.
Seventy percent is not a law of nature. It is the compounding interest on three conversations nobody had in the first hundred days. I have spent thirty years having those conversations for a living — early, in writing, with the board in the room. That is the whole trick, and it is the difference between a hold that compounds and a hold that restarts.
Frequently Asked Questions
What percentage of PE-backed CEOs are replaced during the hold?
Roughly 70% of private-equity-backed CEOs do not survive the hold period, with replacements clustering between months 18 and 24 of ownership — late enough to cost the fund significant runway, early enough that the underlying issues were usually visible and fixable in the first 100 days.
What does replacing a portfolio company CEO cost a fund?
Beyond search fees and severance, the real cost is time: six to nine months of drift, three to six months of search, and roughly six months of new-CEO ramp — a lost year or more on a six-year hold. On a return model driven by EBITDA growth, that lost compounding is frequently the difference between the underwritten multiple and a written-down one.
Why do PE-backed CEOs fail between months 18 and 24?
Because that is when the fund’s exit math first collides with the company’s budget-based trajectory. Small annual variances that budgets hide compound into a visible gap against the underwritten MOIC, and the board — facing push-the-CEO, change-the-plan, or change-the-CEO — chooses the option that feels most like action while a search can still ‘save the deal.’
How can a PE-backed CEO avoid being replaced?
Close three locks in the first 100 days: the Board Lock (the exit number agreed in writing), the Capability Lock (an honest team audit with gaps actioned), and the Team Lock (compensation aligned to the value-creation bridge) — then report monthly against the bridge, by lever, so the board watches the plan rather than being surprised by results.
When should a sponsor actually replace a portfolio CEO?
After the system is fixed, not instead of fixing it. If the number was locked, the team audit was honest, and bridge reporting gave the board real visibility — and the CEO still cannot execute — the seat is genuinely wrong and the change will be fast and clean. Replacing a CEO inside a broken system usually transfers the same failure to the successor.
