Employee Turnover Cost: Beyond the Flat Salary Multiple

You’ve probably heard that losing an employee costs about one and a half times their salary. It’s a handy figure. It fits on a slide, it sounds authoritative, and it’s easy to repeat.

It also falls apart the moment someone asks what’s actually inside it.

Here’s what I’d want a colleague to understand before they quote that number in a budget meeting: the cost of losing someone isn’t a tidy percentage of salary. It’s a stack of separate costs. Recruiting. An empty seat. Onboarding. The slow climb to full productivity. Knowledge handoff. And sometimes the quiet realization that critical context walked out the door with one person. Some of those you can benchmark. Some you have to estimate. A number you can defend is built from the parts, not borrowed from a rule of thumb.

Here’s how to build it.

Which role are you actually costing out?

Start with one specific role, not a company-wide average. Turnover doesn’t cost the same everywhere. A specialist who’s hard to replace and kept years of client history in their head is a very different loss from a role with a deep hiring pool and a clean, documented handoff. Blend them into one average and you’ll learn nothing useful about either.

So before any math, get concrete about this role:

• how many people in it leave in a year;
• their salary and fully loaded compensation;
• how long the seat usually sits empty;
• how much of the work actually goes uncovered while it’s vacant;
• recruiting and onboarding costs;
• how long a replacement takes to reach full speed; and
• how much knowledge and context has to be rebuilt.

Where you have your own numbers, use them. Where you don’t, start from a credible benchmark, label it as a benchmark, and leave it editable. That habit, real data first and clearly marked assumptions second, is what keeps the whole estimate honest.

What does an hour of this person actually cost?

Salary is only part of what an employer pays. Benefits, payroll taxes, and the rest pile real money on top. To get an hourly cost you can plug into everything else, first gross the salary up to fully loaded compensation, then divide by the hours worked in a year.

How much to gross it up? The Bureau of Labor Statistics gives a useful anchor: in March 2026, wages made up 69.9 percent of total employer compensation in private industry, with benefits accounting for the other 30.1 percent (BLS Employer Costs for Employee Compensation). That works out to a loaded-cost multiplier of roughly 1.43. It’s a national average, though, so treat it as a starting point, not a stand-in for your own benefit structure.

What goes into the number, piece by piece?

Now build the stack. Each piece gets counted once, and the one rule to hold onto is to watch for overlap so no dollar gets counted twice.

Recruiting. The cost to source and hire a replacement: departures times your cost per hire. SHRM’s 2025 benchmarking put the average cost per hire at about $5,475 for nonexecutive roles and $35,879 for executive ones (SHRM 2025 benchmarking release). Those are survey averages, not the price you’ll pay. And if your cost-per-hire figure already includes interview and admin time, don’t add that time again separately.

The empty seat. While the role is vacant, some work simply doesn’t get done. Estimate it as the working days the seat sits empty, times the loaded daily cost, times the share of the work that actually goes uncovered. That last part matters: not everything stops. Some work gets absorbed, delayed, or dropped, which is exactly why you don’t charge 100 percent. SHRM’s 2026 data shows a median of 39 calendar days to fill a nonexecutive role (SHRM recruiting benchmarking), though yours may run faster or slower.

Onboarding and training. Direct training costs plus the internal hours spent bringing someone up to speed, valued at the right hourly rate. The new hire’s time and the team’s time may be worth different amounts, so keep them separate.

The ramp to full speed. A new person rarely performs at 100 percent on day one. If they average, say, 60 percent during ramp-up, the cost is the 40 percent gap, not their whole salary for those months. Charge the gap, nothing more.

Knowledge and lost context. This is the piece most estimates miss. Knowledge transfer isn’t just handing over files; APQC describes it as passing along timely, contextual knowledge, including the kind that’s hard to write down (APQC, “What Is Knowledge Transfer?”). It can include the departing person’s handoff time, the replacement rediscovering how things worked, and coworkers reconstructing lost history. One caution from experience: if you’ve already captured that rediscovery inside the ramp-up gap, don’t count it a second time. Only add it where it’s genuinely extra.

What does one piece look like in practice?

Take the empty seat, since it’s the one people argue about most. Imagine a single departure in a role paying $120,000 a year. Grossed up by that 1.43 multiplier, fully loaded compensation is about $171,674, or roughly $660 a day.

Say the seat sits empty for 39 calendar days, which is about 28 working days, and about half the work goes uncovered while it’s gone. That’s 28 days times $660 times 50 percent, or about $9,200.

That’s one piece, not the whole number. Recruiting, onboarding, ramp, and knowledge transfer each still need their own inputs. But notice what showing the piece buys you: an HR or finance partner can push back on that 50 percent, swap in a real time-to-fill, or drop in your actual loaded cost. Nobody can argue with “1.5 times salary.” Anyone can argue with this, and that’s the point.

Why give a range instead of one number?

Some inputs rest on solid benchmarks: recruiting cost, the loaded multiplier, time to fill. Others, like how much work goes uncovered or how long the ramp really takes, are judgment calls. When the shaky inputs move the answer a lot, don’t pretend to a single precise figure. Show a range, and treat the midpoint as a summary, not a promise.

One more distinction worth saying out loud: how the estimate is built is not the same as how accurate it is. It helps to show how much of the number comes from published benchmarks versus your own inputs, but that’s a statement about composition, not confidence. A number built mostly from benchmarks isn’t automatically right, and one built mostly from your own data isn’t automatically wrong.

What should you pressure-test with HR and Finance?

Before this number leaves the room, walk it through with the people who’ll be asked to stand behind it:

• What’s already baked into the cost-per-hire figure, so nothing gets double-counted?
• How much of the departed person’s work genuinely goes uncovered?
• Whose salary should value the internal time spent onboarding?
• Does the ramp cost charge the full salary, or just the productivity gap?
• Are the knowledge-recovery hours truly extra, or already inside the ramp?
• Is any project-disruption figure based on documented labor, or a guess about lost business?
• Which defaults can you now replace with your own data?
• What does the number look like at the low and high ends?

What makes an estimate worth trusting?

A turnover estimate earns trust not by landing on a big, confident number, but by showing its work. A good one makes clear which costs are the usual suspects and which are about lost knowledge, which lean on a benchmark and which came from you. That way it arrives ready to be questioned instead of taken on faith.

Jaspen’s Cost of Employee Turnover Calculator starts you off with visible, editable defaults instead of a blank spreadsheet, so you can lower the uncovered-work share, drop in your real time-to-fill, or replace the loaded cost with your own. It won’t promise that turnover can be prevented, or that every modeled dollar can be won back. What it gives you is a figure you can defend, the kind you can carry into a budget conversation without having to explain where it came from.