What Counts as Attrition? An HR Manager Answers the Tricky Questions

Let's Start With the Obvious: What Actually Is Attrition?

I've been an HR manager for eleven years, and I still get this question in some form every quarter — usually from a department head who's just lost two people and wants to know whether they should be worried. The honest answer? It depends on which two people, why they left, and what the rest of the team looks like right now.

At its core, attrition is the reduction in headcount that happens when employees leave and you don't replace them immediately — or at all. That last part is the bit most people overlook. If someone quits and you backfill the role within a week, many HR teams would call that turnover, not attrition. The terms get used interchangeably in casual conversation, but they carry meaningfully different implications when you're analyzing workforce data.

So let's work through the questions I actually get asked, because that's more useful than yet another glossary definition.

Q: Someone resigned. Is that always voluntary attrition?

Almost always, yes — but not always. Here's where it gets nuanced.

If an employee hands in a resignation letter because they found a better offer, that's clean voluntary attrition. But what if they resigned because their manager was creating a hostile environment and HR hadn't resolved it after months of complaints? Technically voluntary. Functionally? That's a management failure driving someone out, and I'd want that flagged differently in my analysis.

Some organizations use a subcategory called push attrition — exits that are technically voluntary but driven by internal dysfunction rather than external opportunity. It's not a universally adopted term, but it's useful. When you see a spike in resignations from one team, one location, or one reporting line, you should ask whether employees were pulled toward something better or pushed away from something worse.

The exit interview is your best tool here, assuming people trust it enough to be honest. Anonymous post-exit surveys sent 30 days after someone leaves often get more candid responses than the formal conversation on their last day.

Q: What about layoffs and terminations — do those count as attrition?

This is where terminology gets genuinely contentious. Here's how I think about it:

• Involuntary attrition includes terminations for performance or conduct, as well as layoffs. The employee didn't choose to leave — the organization made that decision.
• Voluntary attrition is when the employee initiates the departure.
• Some frameworks treat attrition as voluntary-only and use "turnover" as the umbrella term for all separations. Others lump everything together.

My recommendation: be consistent within your own organization and explicit about your definitions whenever you're sharing numbers with leadership. If your CFO thinks "attrition rate" means voluntary-only and you're reporting combined figures, you're going to have a confusing conversation.

For what it's worth, when I build dashboards, I always break it out: voluntary, involuntary (performance), involuntary (restructuring/layoff), and retirement. Four buckets. Each one tells a different story and calls for a different response.

Q: What is "regrettable attrition" and who decides whether a departure is regrettable?

This is my favorite question because the answer makes some executives uncomfortable.

Regrettable attrition refers to departures where the organization genuinely wanted to retain the person. They were a strong performer, they had institutional knowledge, they were on a leadership track — something. The loss hurts.

Non-regrettable attrition is the inverse: performance-managed exits, people who were in roles they'd outgrown poorly, or situations where the departure actually opened up space for someone better suited.

Here's the uncomfortable part: who decides? Ideally, this classification happens through a structured process — performance ratings, flight risk assessments, succession planning data — not just a manager's gut feeling about whether they liked the person. I've seen "non-regrettable" used as a post-hoc rationalization for losing someone who should have been retained with a better counter-offer or a different manager. That's not honest analysis; that's covering your tracks.

A clean process looks like this: before your quarterly talent review, HR and managers pre-classify current employees into retain/develop/manage-out tiers. When someone leaves, you check which tier they were in. If someone in your "retain" tier walks out the door, that's regrettable — full stop, no revision after the fact.

Q: How do I calculate attrition rate, and does the formula actually matter?

Yes, the formula matters — especially if your headcount fluctuates significantly during the year.

The basic formula most people use:

Attrition Rate = (Number of Separations ÷ Average Headcount) × 100

Average headcount is typically calculated as: (headcount at start of period + headcount at end of period) ÷ 2.

For a stable organization, this works fine. But if you hired 200 people in Q1 and then had significant attrition in Q3, your average headcount calculation can produce numbers that feel misleading. Some organizations use a rolling 12-month average headcount instead, which smooths out those hiring surges.

A few other things that skew attrition numbers:

• Contract and contingent workers — are they in your denominator? If yes, their departure patterns may look very different from permanent employees, and blending them muddles your analysis.
• Internal transfers — if someone moves from one business unit to another, does that show up as attrition for the sending team? For some reporting purposes, yes. Make sure your leadership knows whether the number they're looking at is company-wide or department-level.
• Parental leave or LOA endings — if someone on extended leave decides not to return, that's an exit. Count it honestly.

Q: Our attrition rate looks fine on paper but the team feels hollowed out. What's going on?

This is the question that separates HR managers who understand their workforce from those who just report numbers.

Aggregate attrition rates hide distribution problems. A 12% annual attrition rate sounds manageable — until you find out that 9 of those percentage points came from your senior engineers and technical leads, while the junior roles have near-zero turnover because there's nowhere for those people to go externally yet.

The dimensions you should always cut your attrition data by:

1. Tenure band — are you losing people in their first 18 months (onboarding failure) or at the 4–6 year mark (career plateau)?
2. Performance tier — are your A-players leaving at higher rates than your average performers? If yes, that's an urgent signal.
3. Manager — some managers have 5% attrition, some have 40%. This is rarely discussed openly enough.
4. Role level — individual contributors vs. managers vs. senior leaders often have completely different attrition drivers.
5. Geography or function — a market with a hot job market for your specific talent type will always show higher voluntary attrition; factor that in before drawing conclusions.

When a leadership team tells me the overall number "looks fine," I pull the distribution. More often than not, there's a pocket of serious damage hiding inside a decent average.

Q: Is there a "healthy" attrition rate, or is that a myth?

Partially a myth. Here's my actual view after a decade of this work.

There's no universal healthy rate. A fast-scaling startup might run 25% annual voluntary attrition and be completely fine — their business model expects rapid role evolution and they've built systems to absorb it. A hospital ICU running 25% nursing attrition is in a patient safety crisis.

What I'd say instead: every organization should know its own baseline and benchmark against its own history first, then against direct industry competitors second. Generic benchmarks ("X% is the industry average") are a starting point, not a verdict.

Also — and I say this gently — zero attrition is not the goal. Some attrition is healthy. It refreshes perspective, creates promotion opportunities, and occasionally exits people who weren't growing with the organization. The goal is attrition that's predictable, understood, and concentrated in places that don't damage your core capability.

Q: Any final advice for HR teams trying to read their attrition numbers more honestly?

Three things I'd leave you with:

First, don't let the classification happen retroactively or politically. Set your definitions and your talent tiers before people leave, so the analysis reflects reality rather than post-exit spin.

Second, pair your quantitative attrition data with qualitative signals — stay interviews, engagement pulse scores, Glassdoor patterns. The number tells you something is happening. The conversations tell you why.

Third, report attrition to leadership in a way that demands a response. "Our voluntary attrition is 18%" is a data point. "Our voluntary attrition among high performers is 28%, concentrated in two business units with the same skip-level manager" is a conversation that leads somewhere.

Attrition isn't just an HR metric. It's a leading indicator of organizational health. The teams that treat it seriously — and read it honestly — tend to catch the problems that everyone else notices six months too late.