Regretted attrition counts only the departures you wanted to prevent — the people you would have kept if you could. Total turnover can look perfectly healthy while your best people quietly leave. Track the regretted number separately and you are measuring the loss that actually costs you.
Every departure, counted the same — the low performer you were ready to let go and the top performer you would have fought to keep. It tells you how many people left, not whether losing them hurt.
Only the departures you did not want — the people you would have kept if you could. It isolates the exits that actually cost you, which makes it the retention number worth managing.
Regretted attrition is the portion of your turnover made up of people you wanted to keep. Everything else — the low performer you were ready to part with, the bad-fit hire who was never going to work out, the person who left a role you were about to redesign anyway — is not regretted. Both kinds of exit show up in the same headline turnover figure, and that is exactly the problem. One number is being asked to describe two completely different events.
This is how a company ends up with “healthy” total turnover while it bleeds its best people. If a chunk of your departures are exits you did not mind, your overall rate can sit comfortably in a normal range while the regretted slice climbs underneath it. The average looks fine. The reality is that the people carrying your hardest work are the ones walking out, and the blended number gives you no way to see it. You cannot manage what you refuse to separate — and the research says as much as 75% of that loss is preventable.
Splitting the two changes what you pay attention to. A rising regretted rate is a signal that something in the work, the manager, or the path forward is pushing out the people you most need — often the same people who show up in guides on quiet quitting and burned-out top performers before they resign. A flat or falling regretted rate, even alongside ordinary total turnover, means your retention is working where it counts.
Tracking regretted attrition does not need a new system — it needs one honest flag at every exit and the discipline to report it on its own. The practice looks like a checklist:
The exit flag is honest, but it is late. By the time you mark a departure “regretted,” the decision is already made and the person is already gone. That makes the metric a good scoreboard and a poor alarm. To act instead of tally, you have to see the shift before the resignation.
Behavioral commitment signals are what surface an at-risk keeper early. When someone who used to take initiative stops reaching beyond their role, when the grit that showed up on hard problems goes quiet, when curiosity narrows to just getting through the day — those are the changes that precede a regretted exit by weeks. Reading them is the difference between measuring commitment as it moves and finding out at the exit interview. It is also the groundwork for a real plan to retain your top performers while there is still time to act.
Regretted attrition is the share of departures made up of people you wanted to keep. Instead of counting everyone who leaves, it isolates the exits you would have prevented if you could — your strong performers and hard-to-replace contributors. Because it strips out the departures you did not mind, it is a truer measure of whether you are holding on to your best people.
At each exit, flag the departure as regretted or not regretted — would you have re-hired this person for the same role today? Then divide the number of regretted departures over a period by your average headcount for that period. Keep it as its own rate, separate from total turnover, so the two numbers can be compared rather than blended.
Total turnover lumps together the exits you welcomed and the ones that hurt, so it can look healthy while you are losing exactly the people you most needed. Regretted attrition counts only the keepers, so it moves when your real retention problem moves. It is the number that tells you whether your talent base is getting stronger or quietly draining away.
There is no universal benchmark, and any single figure treated as a hard target tends to mislead. As a general rule of thumb, you want regretted attrition to be a small and shrinking share of your total turnover — most of the people leaving should be departures you were fine with. The direction of the number over time matters more than hitting a specific percentage.
The exit flag is a lagging signal — by the time you record it, the person is already gone. To get ahead of it, watch behavioral commitment signals on the people you most want to keep: fading initiative, grit that used to show up on hard problems going quiet, curiosity narrowing. Those shifts surface at-risk keepers weeks before they resign, while you can still act.
Book a call to see how Prove surfaces at-risk keepers from real behavior — before the regretted exit, not after.