Here's the arithmetic and how to run it on your own HRIS data.
Somebody good left last quarter. The file codes it as voluntary.
That is the last time anyone looked at it. It reached a total turnover of twelve percent. This figure went onto a slide for the board. They noted that twelve percent matches the industry benchmark. Nobody in the room was wrong. The rate is normal. The benchmark is real. The most costly issue in your organization went unnoticed in that meeting. An aggregate averages costs, and averages can hide them.
Here is the number that was in the room but not on the slide.
The Mechanism
Most organizations experience differential turnover. This means people leave at different rates based on their groups. Not everyone leaves at twelve percent. Some groups in your company are at eight. Another is at sixteen. The consolidated is the only number anyone looks at, and it is the one number that cannot show you this.
The spread is the cost, not the rate.
Run it. The replacement cost for a professional employee is about $220,000. This is roughly double the $110,000 salary, which is on the higher end of estimates. That figure is an assumption, not a measurement. You should replace it with your own. The published estimates range is 30–50 percent of salary for entry-level roles. For professional roles, it goes from 100–150 percent. Specialized positions can reach 200 percent or more. See this SHRM article for more context. Use whatever your finance team will defend.
Now take the spread. In most organizations, the difference in attrition rates is about ten percentage points. The group with the lowest attrition often sees much better retention than the one with the highest. In a company of a thousand people, this means around seventy-five to ninety extra departures each year. These are departures you could avoid if every group left at the same rate as your best-retained group.
Multiply. Seventy-five departures at $220,000 cost $5.5 million a year or $16.5M over three years in replacement expenses. This doesn’t even include what’s lost when knowledge walks out the door.
The number scales with both inputs. Run it with your own salaries and your own spread, and you may land somewhere else. But for a company of any size, the order of magnitude holds, and it is not a rounding error. It is a line item that would trigger an immediate review in any other part of the business.
Why I Went Looking for This
I got an autism diagnosis at fifty-five. Before that, I spent forty years leading. I was in the Navy, and later worked at Dell, Tech Data, and then started two of my own companies. The diagnosis helped, but a number I found later shifted my focus.
In 2024/25, 31.4% of working-age people in the UK whose main long-term health condition is autism were in employment (Department for Work and Pensions, The Employment of Disabled People 2025).
People with in-demand technical skills are sitting at home. They struggle to pass interviews that check eye contact and conversational skills, even when trying to seem relaxed.
I posted on LinkedIn offering to help anyone who wanted to prepare for one. The response came from all over the US and Europe: engineers, analysts, researchers, and writers. Every conversation showed the same thing. I want to be clear: the insight came from them, not me. I recognized its shape. After thirty years on the hiring side, I saw the same filter at work. Yet, I didn’t know what I was looking at.
What I was seeing was not an autism problem. It was the ordinary problem, running at its most extreme and thus its most visible. Systems that exclude autistic candidates filter out introverts from leadership roles in the same way. They also remove concrete thinkers from strategy jobs. People whose processing style doesn't match the template get excluded from decision-making. Same mechanism, different degree. Most of your people pay that tax at a low rate. It's enough for them to survive but high enough to cost you something you haven’t measured.
Some of them leave. That part shows up in your HRIS. Many stay and contribute less than they could, and that part shows up nowhere at all.
Why Your Exit Interviews Didn't Catch It
"Seeking new opportunities." "Better compensation elsewhere." "Time for a change."
I've written those. When you leave a place that has worn you out, you don’t write that on a form with your name. You are polite, you shake hands, and you leave with your reference intact. Everyone involved understands the transaction. The exit interview is the final chance for a departing employee to share their thoughts honestly. It's the main tool many organizations use to learn why people leave.
This is not an HR failure; it is a structural one. The instrument asks for information that the person may want to keep secret.
Your HRIS has no such problem. It doesn't rely on anyone's candor. It records who left, when, and from where, and the pattern sits in it whether anyone was willing to name it out loud.
Run it Yourself
This takes an analyst an afternoon.
-
Pull two years of separations from your HRIS.
Calculate the turnover rate for each group you can analyze:
-
Function
-
Tenure band
-
Level
-
Manager
-
Demographic
-
Use whatever dimensions your data supports.
-
Take the group with the lowest rate; that is your baseline. It shows what your organization can really keep. This is not just a guess.
-
For every other group, calculate the excess departures against that baseline.
-
Multiply excess departures by your replacement cost.
The manager cut is usually the one that produces the longest silence. Company-wide averages can mask individual managers. For instance, a 3.8 average psychological safety score may hide a 2.6 for one team and a 4.2 for another. The disaggregated view is where anything actionable lives.
Here’s what to expect when the number lands. I've noticed it too: the silence in the room feels unique after a diversity presentation. It is the silence of a CFO doing arithmetic.
What You Could Notice This Week
You don't need the full analysis to start. You need one question, asked of your HRIS: what is the spread between my best-retained group and my worst?
If it's two points, you have a smaller problem than most companies, and you should verify the data. If it's eight, you have the standard version, and you now know roughly what it costs.
You’ll learn something your turnover rate can’t show you. Plus, you’ll know it before your next board meeting, not after a resignation.
Run your differential turnover number → go to calculator
Enter your headcount, your replacement cost assumption, and your attrition spread. It returns the annual figure and a one-page summary you can take to your CFO.
See resources with more information.
Peter Allen Mann is the author of The Perception Revolution (Fast Company Press) and The Projection Fallacy (April 2027). He spent four years as a Navy surface warfare officer. Then, he worked for a decade at Dell and Tech Data. After that, he spent two decades building companies. He maintained about 95 percent voluntary retention in all them.