Total attrition is the sum of very different things. Reporting the total only is almost always misleading.

Regrettable attrition: employees you wanted to keep left (voluntary, performer tier ‘meets’ or better).
Non-regrettable attrition: employees you were fine seeing leave (involuntary terminations for performance, PIP-driven exits, structural reductions).
Retirement / structural: retirements, end-of-term contracts, international relocations — not usually counted as either.

Why the split matters:

An 18% total with 14% regrettable is a crisis.
An 18% total with 6% regrettable and 12% non-regrettable is a deliberate performance-management cycle — leadership may even be happy.
The HR interventions are completely different. Regrettable attrition argues for retention incentives, manager coaching, compensation review. Non-regrettable attrition argues for hiring bar adjustments or performance-management calibration.

What a good people-analytics report includes:

Regrettable vs non-regrettable (rate, not count).
Segment by tenure band (0–1y, 1–3y, 3–5y, 5y+). High 0–1y attrition signals onboarding or mis-hire problems.
Segment by function and level. Engineering senior-level regrettable is very different from support entry-level regrettable.
Segment by manager. If 5 of 10 regrettable departures came from the same manager, that’s the story.
Segment by demographic, carefully, behind an HRBP gate. Gaps by gender, race, or age tell a culture story and deserve leadership attention — but the data is sensitive.

What to tell the VP: “The 18% number hides whether you have a retention problem or a standards-raising cycle. I’d split it into regrettable / non-regrettable, by tenure band and function, before we pull any intervention lever.”

What’s the difference between regrettable and non-regrettable attrition?

Draft your answer

Stuck? Peek at a hint

More people analytics questions