The 90-day regret window on accepted counter-offers

A senior person hands in a resignation, the company counters, and they accept. Everyone treats the file as closed. Retention saved. The reality is narrower than that: the counter-offer settled the compensation gap and nothing else. The compensation gap is rarely the whole reason a Director or VP starts interviewing in the first place. It's the reason that fits on a resignation letter.

Underneath the number sits the driver that actually moved them: a scope that stopped growing, a manager they no longer learn from, a mandate that has narrowed while the title stayed the same. A raise doesn't touch any of that. So the person stays, the pay resets, and the original grievance goes quiet for a while because the money bought a period of relief. That relief is temporary.

That's why acceptance is a poor predictor of retention. Published estimates on counter-offer failure vary widely and none of them are clean, so we won't quote a single figure as fact. What isn't in dispute among people who recruit at this level is the shape: a meaningful share of accepted counters unwind within the year, and they unwind on a schedule. The money buys time. The clock starts the day they say yes.

There's a rhythm to how an accepted counter decays. The first few weeks are calm. The raise feels good, the counter felt flattering, and the effort of a job search has just been spent, so the path of least resistance is to stay and enjoy the reset. This is the honeymoon, and it's real, and it's misleading, because it tells the manager the problem is solved.

Then the original driver comes back into view. The scope that pushed them to look is still the scope. The manager they didn't rate is still the manager. Around the end of the first quarter, the raise has been fully absorbed into their baseline, it's now just their salary, and the thing money was papering over is visible again at full size. This is the regret window: the stretch, roughly the first 90 days, where the person quietly re-litigates the decision to stay.

The dangerous part is that the regret window is silent. A person re-opening the question doesn't announce it. They aren't unhappy in a way that shows up in a one-to-one, because they made the choice themselves and are giving it a fair run. So the manager, reading calm on the surface, banks the retention as won at exactly the moment the decision is being reconsidered underneath it.

By the time someone resigns a second time, the decision was made weeks earlier. The useful signals sit ahead of that, inside the window, and they're behavioural rather than stated. A manager watching for them can act while acting still changes the outcome.

The first is that the scope hasn't moved. The counter promised, implicitly or out loud, that things would be different. If the same person is doing the same work with the same mandate 90 days later, the promise has quietly failed, and they know it before you do. Unchanged scope after a counter is the single clearest tell that the retention is decaying.

The second is that the raise has stopped reading as recognition. In week 2, the number felt like the company valuing them. By month 3, it's the going rate for the job they already had, and any warmth attached to it has evaporated. When a person stops mentioning the raise entirely, it has become baseline, and its retaining power is spent.

The third is external re-engagement. They start taking recruiter calls they were ignoring a month ago. Their profile gets a quiet update. A reference request lands from a peer. None of this is visible on an org chart, but it's exactly the pattern that precedes a second, final resignation, and it tends to cluster right at the end of the first quarter.

There's a reason managers keep misreading this. The act of accepting a counter changes how the person behaves in a way that hides the underlying problem. Having chosen to stay, they invest in that choice. They're motivated to make it work, at least for a while, which produces a stretch of genuine effort that looks exactly like a resolved situation.

The reset of the money also does real short-term work. A pay bump lands as recognition in the moment, and recognition is a legitimate retention lever, briefly. The problem is duration. Recognition delivered as a one-time number decays fast, because the person adapts to the new figure and it stops signalling anything. What began as “they fought to keep me” becomes “this is what the job pays,” and the emotional return on the raise is gone well inside the quarter.

Meanwhile the driver that money never addressed, the flat scope, the stalled growth, the manager relationship, is untouched and patient. It doesn't decay. It waits. So the two curves cross: the retaining power of the raise falling, the original grievance holding steady, and somewhere around the end of the first quarter the grievance is once again the larger force. That crossing point is the regret window, and it's structural, which is why it recurs.

The window is short, but it's the one stretch where intervention still works, because the person hasn't yet privately closed the decision. A second raise inside a quarter reads as panic and confirms that the company only responds to a threat. The move is to spend the window resolving the thing the money never touched.

That starts with naming it. A manager who says, plainly, “the raise doesn't fix why you were looking, so let's talk about that,” does something the counter-offer itself never did: treats the actual driver as real. Most retained hires are never asked this, because the resignation-and-counter is treated as a closed transaction. Reopening it, deliberately, in week 3 or 4, is the single highest-value conversation available.

Then the scope has to move visibly. Whatever was promised in the heat of the counter needs a concrete change the person can see inside the window: a mandate that widens, a stalled project unblocked, a growth path put in writing with a date on it. 90 days of unchanged work will undo any words spoken during the counter. And the honest last step is triage: some counters were always going to fail because the real driver was the manager or a ceiling the company can't lift, and recognising that early lets you plan the succession ahead of the second resignation.