Employee turnover cost calculator

When a $65,000 employee leaves, replacing them costs $26,410 in this example, 40.6% of salary: $1,500 of separation, $5,800 of recruiting, $7,031 for the empty seat, $2,600 of onboarding and $9,479 for the months in which the new hire is paid but not yet fully productive. With 120 employees and 22% turnover, that is 26 departures and $697,235 a year. Published estimates run from 21% of salary to 200% or more, which is a tenfold range.

This calculator builds the cost of one departure from your own numbers, component by component, and puts it beside the published figures as a sanity check. It then values what a lower turnover rate is worth and the most a retention effort can spend before it stops paying.

Who reaches for this

The calculator asks for the salary of a role, the payroll load on top of it, and the cost of each stage of replacing someone. Loaded cost is the salary plus payroll taxes and benefits. The daily and monthly figures come from it: divide by 260 working days or 12 months. Each stage then has its own inputs, and the total is their sum.

Two calculations follow. The company-level view multiplies the cost per departure by the number of departures a year, which is headcount times the turnover rate. The retention view multiplies the departures a program prevents by the same cost, and subtracts the program’s price. The true cost of an employee calculator covers the loaded cost that goes into the vacancy and ramp-up lines, and the revenue per employee calculator helps judge how much output a role is worth.

Search for the cost of turnover and the same few numbers appear on every page. Tracing them is harder than it should be, and the result is that they are far less solid than they look. Here is what is claimed, who is said to have claimed it, and what I could confirm.

The 21% figure is the one with a traceable method. It is the median of case studies that were mostly about jobs paying under $50,000, and it leaves out the roles where the cost is greatest. The 50% to 200% range is more often quoted and less documented. Different pages credit it to different organizations, and one page splits it across them in a way that does not match the others. A range cited by several sites is not the same as a range that has been checked.

Why they differ so much

None of that makes the published numbers useless. It makes them a range to check yours against. If your build-up gives 15% of salary for a warehouse role and 110% for an engineer, both fall inside the published range for their kind of role, and you know why. If you get 400%, look for a component that is counted twice.

Six components cover most of the cost. The first three are events and the last three are periods of time.

The vacancy and ramp-up lines both use the loaded cost of the role as the value of the work, which is a choice worth understanding. A role that produces more than it costs, such as a salesperson, is worth more than its loaded cost, and using the cost understates the loss. A role that produces less than it costs is worth less. The convention is conservative for revenue roles and generous for support roles. For a role where you can estimate the revenue effect, use the other costs line. Recruiting cost is built up in detail on the cost per hire calculator.

Where to find your own numbers

Most of the inputs are already somewhere in the business. Separation costs come from the last few exits: what was paid out in accrued leave and how many hours HR and the manager spent. Recruiting costs are on invoices and in the applicant tracking system, and interview hours can be estimated from calendars. Time to fill is the gap between the departure and the start date of the replacement, across recent hires in the same role.

The two that need judgment are the ramp-up and the share of vacant work that goes uncovered. Ask managers how long it takes someone in the role to reach full speed and how far short of it they are along the way, and ask what actually happens to the work in the meantime. Ranges are fine. Run the calculator at the low and high ends and report both, since a range that is honest is worth more than a single figure that is precise.

The example is a company with 120 employees, a mid-level role paying $65,000 with a 25% payroll load, and a loaded cost of $81,250. It takes 45 working days to fill the seat, and half of the work goes uncovered meanwhile. The new hire needs four months to reach full productivity and averages a 35% shortfall over that time.

Across the company, 22% turnover of 120 people is 26.4 departures a year, and 26.4 × $26,410 is $697,235. Loaded payroll is 120 × $81,250 = $9.75 million, so turnover costs the company 7.2% of its payroll every year. The figure is a reason to ask what the company spends on retention now, and what it would take to move the rate.

The $26,410 sits above the 21% median from the case-study review, which would be $13,650 for this salary, and below the low end of the widely quoted range, which would be $32,500. That is a plausible place for a mid-level role with a four-month ramp. The published figures do not tell you whether it is right. They tell you it is not absurd.

The cost per departure is only half the calculation. The other half is how often departures happen. Turnover rate is the number of employees who leave in a period divided by the average headcount over that period. A company that lost 26 people with an average of 120 employees has a rate of 26 ÷ 120 = 21.7%.

Use the average headcount, not the starting or ending count. A company that grew from 110 to 130 people has an average near 120, and dividing by the starting figure overstates the rate while dividing by the ending figure understates it. The same point about averaging headcount appears in the revenue per employee calculator. For a shorter period, annualize carefully: a monthly rate of 1.8% is about 21.6% a year, and a single busy month is a poor basis for a yearly estimate.

Split it before you cost it

The total hides differences that change the cost. Voluntary turnover, where people choose to leave, is the part a company can influence, and involuntary turnover, dismissals and layoffs, has different costs and causes. First-year turnover is worth its own line, since it wastes the whole investment in hiring and onboarding. And departures of people the company would have liked to keep cost more than those it accepts. A calculation by group, such as by role or by tenure, is more useful than a single average.

In the example, the visible hiring costs, meaning separation, recruiting and onboarding, are 37.5% of the total. The time lost to the empty seat and the ramp, vacancy and ramp-up together, is 62.5%. A company that measures only what shows up on invoices sees the smaller part.

That points to different fixes. Cutting ad spend saves part of a line that is 22% of the cost. Shortening the vacancy and the ramp addresses more than half of it. Each one is sensitive to the assumptions, which is worth testing.

Halving the ramp saves $4,739 a departure, and cutting the vacancy from 45 days to 20 saves $3,906. Faster hiring and better onboarding pay back on every departure that still happens, even if turnover does not change.

A lower turnover rate is worth the cost per departure times the departures avoided. With 120 employees, each point of turnover is 1.2 people, and each is worth $26,410, so each point is worth $31,692 a year. That is the number to compare with what a retention effort costs.

An $80,000 effort breaks even at 3.03 departures avoided, or 2.5 points of turnover. Any reduction beyond that is net saving. The question for any raise, benefit or program is whether it can plausibly do that. A change that costs $500 per employee across 120 people is $60,000, and it pays for itself if it prevents 2.3 departures. The figure to hold the effort to is the departures it must prevent.

Be careful with the claim. A program that coincides with a fall in turnover has not proved that it caused the fall, and the savings should be estimated from what is likely, not from the best case. Count the effort at its full cost, including manager time, and treat the result as a range. The value is in knowing what break-even looks like before you spend.

Is a raise cheaper than a departure?

The same arithmetic tests a pay decision for a single employee. A $3,000 raise on a role with a 25% payroll load costs $3,750 a year. Against a $26,410 cost per departure, it pays for itself if it lowers the chance that this person leaves in the year by 3,750 ÷ 26,410 = 14.2 percentage points. If you believe the raise moves the odds by more, it is the cheaper option. If the person is likely to stay either way, it is not.

Two cautions apply. The raise is permanent and compounds with later increases, while the cost of a departure is incurred once. And a raise for one person can raise expectations for others. Use the break-even probability as a way to state the bet in plain terms before making it.

One average hides very different costs. The same six components, run for three roles with inputs that scale with seniority, give quite different results.

Three things show up. The share of salary rises with seniority, because senior roles take longer to fill and ramp up and carry more that is hard to replace. The frequent, cheap departures still add up: 21 entry-level departures cost $120,976. And the yearly total, $1.18 million across the three groups, is dominated by the middle group, where both the rate and the cost are high. That is where an improvement in the rate pays the most.

The inputs for these roles are illustrations chosen to show the pattern, and yours will differ. Fill in each role from what you know: the agency fees you pay for senior hires, the weeks they stay empty, and how long it takes someone in that job to reach full speed.

In a small business

The percentages do not shrink in a small company, and the effects concentrate. Take a six-person firm with a $48,000 role, a 20% payroll load, $500 of separation, $600 of job-board costs and 25 hours of owner time at $45, 30 empty days with 60% of the work uncovered, $300 of onboarding and 20 hours of training time, and a three-month ramp at a 30% shortfall. The cost of one departure is about $11,700, or 24% of salary, and it arrives as owner hours and missed work more than as invoices.

In a team that small, one resignation can be a sixth of the workforce, so the disruption is larger than the arithmetic. It is also where the calculation has the most to teach: an owner who counts their own hours at a real rate usually finds that hiring is the most expensive job in the business.

Putting the number in a budget

Departures a year, times the cost of each, is the exposure. Only part of it is cash. In the example, separation, recruiting and onboarding come to $9,900 a departure, so 26.4 departures need about $261,000 of spending. The other $436,000 is lost time, which does not appear as an expense line but shows up as missed output, overtime and slower delivery. A budget that plans for the cash and ignores the time will look fine and still underperform.

Some costs of turnover are real and resist measurement. Knowledge leaves with the person: how a system was set up, why a decision was made, what a client prefers. The colleagues who cover the gap absorb extra work, and that can raise their own risk of leaving. A departure can also unsettle a team, and in a client-facing role it can put a relationship at risk.

Two approaches keep this honest. For revenue at risk, estimate the margin on business that is likely to be delayed or lost because of the departure, and enter it under other costs. A salesperson who leaves an account for two months may put $30,000 of revenue at risk, and at a 50% margin that is a $15,000 cost the base method misses. For everything else, list the effect without pricing it. A calculation that says “$26,410, plus effects on the team that are not counted” is more defensible than one that invents a number for morale.

The opposite mistake is to count nothing. Companies that price only the invoices see a cost of a few thousand dollars per departure and conclude that turnover is cheap. The truth is usually in between: a defensible number built from components you can explain.

See what a hire costs to keep with the true cost of an employee calculator, or compare hiring with contracting using the employee vs. contractor calculator.