Preventable turnover
Most Turnover Is Preventable. So Why Do Companies Keep Losing People?
If most departures could be stopped, the real question is why they keep happening anyway. The answer is in the gap between seeing it and acting on it.
June 6, 2026 · 5 min read
Exit interviews keep producing the same uncomfortable result. The Work Institute, which analyzes exit interview data, has reported that roughly three in four voluntary departures were preventable by the employer. Sit with that for a moment. It says most of the people who walked out were not lost to a relocation or a genuine retirement. They were lost to something the company could have influenced and did not.
So the obvious question is not why people leave. It is why departures a company could have stopped keep happening anyway. The answer is not that companies do not care, because most of them spend real money trying to retain people. Our view is that the answer is structural. Noticing that one particular person is at risk and doing something about that person are two different acts, performed by different people at different times, and most retention programs do the first in bulk and never connect it to the second.
Preventable does not mean prevented
The word preventable does a lot of quiet work. It means the reason a person left was something the organization could have influenced. Not a spouse’s job relocation or a genuine retirement, but the things that show up over and over in exit data: a manager who did not listen, work that stopped feeling meaningful, no visible path forward, recognition that never came.
Those causes are addressable. A different conversation, at the right time, changes the outcome for a real share of people. That is what an exit interview finding of this kind is really measuring, how many departures trace back to something inside the company’s control rather than to life events nobody could have negotiated with.
So the gap is not knowledge. Companies broadly know why people leave, and it is not a gap in effort either. The gap is between knowing it in general and seeing it in a specific person while there is still time to act.
Where the standard playbook breaks
Walk through what most companies actually do about retention and the failure point becomes clear.
Every one of these operates on the group. Turnover risk does not live in the group. It lives in specific people, for specific reasons, on specific timelines. A strategy aimed at the average will always miss the individual who is actually leaving.
There is a second pattern in that list. Each item either measures something or does something, and none of them do both about the same person.
The annual engagement survey.
It tells them how the organization felt, on average, months ago, among the people willing to answer honestly. It cannot tell a manager that the person two desks over is three weeks from deciding to leave.
Compensation benchmarking.
Useful, but blunt. It treats everyone the same and misses that the person at risk is rarely leaving over pay alone.
Recognition programs and wellness perks.
These address real drivers, but at the level of the whole population, not the individual whose specific unmet need is going unspoken.
Manager training.
It happens in a workshop twice a year and is forgotten by the following Monday.
The timing problem underneath everything
There is a second reason preventable turnover keeps happening, and it is about when, not what.
Most voluntary departures are preceded by a period of withdrawal. The person has not handed in notice, but they have started to disengage and, often, to look. That window can last weeks or months. During it, the outcome is genuinely open. A conversation that takes the person’s situation seriously can pull them back.
The trouble is that the standard tools are all lagging indicators. The engagement survey reports the past. The exit interview happens after the decision is final, which is the cruelest timing of all, learning exactly why someone left on the day they leave. By the time the system registers a problem, the window has usually closed.
Timing is also where the split between seeing and acting does its damage. HR owns the instruments, the manager owns the relationship, and what the instruments notice reaches the manager late, in aggregate, and stripped of the only thing that would make it usable, a name and what to say to them. Two capable parties, both doing their jobs, and the risk falls into the seam between them.
Prevention requires catching the signal inside the window and putting it in the right hands while the window is still open. That means looking at individuals, regularly, and comparing what the employee experiences against what their manager believes, because the gap between those two views is one of the earliest and most reliable signals there is.
What actually moves the number
If the failure is one of individual sight, timing, and handoff, then the fix follows from the diagnosis.
It means reading people one at a time, not just surveying them in bulk. It means doing it on a regular cadence, so a person who was settled six months ago and is drifting now gets caught this quarter rather than at their exit interview. It means putting the insight where the leverage is, in the hands of the manager who can actually have the conversation, not just in an HR dashboard. And it means handing that manager not a risk score but a plan, the specific thing to say and do, before the notice lands.
That last step is the one most retention work skips, and we would argue it decides everything. A risk score tells a manager to worry. It does not tell them what to do on Tuesday. Detection that ends in a number relocates the problem rather than solving it.
That is the difference between knowing turnover is preventable and actually preventing it. The first is a statistic. The second is a practice.
Companies keep losing people they could have kept because they are equipped to understand turnover in the aggregate, not to see it coming in the one person about to walk, and because seeing it and acting on it live in different hands. Close that seam and the preventable share stops being a number in a report and starts being people who stayed.
Anchor reads one employee at a time and hands their manager a clear picture and a plan, in time to act.