A resignation is experienced as a staffing problem. It is also a financial event, and it is one of the few costs an organization carries repeatedly without ever putting a number on.
The number matters for one specific reason: retention costs money too. Higher pay, better supervision, smaller caseloads, actual training — all of it appears in a budget as an increase, and it gets weighed against a turnover cost that has never been calculated. That is not a fair comparison. It is a known cost against an unknown one, and known costs lose.
Four buckets, and the invisible one
Coverage while the post is vacant. Overtime, agency or relief cover, and the work absorbed by colleagues. The absorbed work is real even though it never appears as a line — it shows up later as someone else's resignation.
Recruitment. Advertising and any agency fee, plus the hours: shortlisting, interviewing, references, and the clearance process, counted at the loaded cost of the people doing it. Managers' time is the largest part and the part most often left out.
Onboarding. Pre-service training, the induction period, the supervisor time that goes into it, and the equipment and licenses issued.
The ramp. The months between starting and working at full effectiveness. This is the invisible bucket, it is usually the largest, and almost nobody counts it — because nothing is billed for it and nobody writes it down.
The ramp is not a criticism of new staff. In case-carrying work a person needs to learn the systems, the standards, the local network, and the specific families or clients on the caseload. Until then they are producing less, taking more supervision, and generating more rework. That is a cost, and it lasts long after the vacancy is filled.
Do not go looking for a benchmark
There is a strong temptation to search for a percentage — some share of salary that turnover is said to cost. Resist it.
Published multipliers come from studies of particular sectors, particular roles and particular labor markets, and applying one to your organization produces a number that is precise, unarguable-looking, and not about you. A finance committee is right to reject it, and once they have rejected the number they will reject the argument attached to it.
Your own arithmetic, with assumptions written down beside it, is worth more than any borrowed figure — even when it is rougher. Especially when it is rougher, because you can defend every line.
Where the estimate should be conservative
Anyone building this number is arguing for something, and that means the estimate has to be built to survive scrutiny rather than to win.
Use loaded cost, not salary, for every hour of anyone's time. Count only the hours you can describe. Where you are unsure of a duration, take the shorter one. Where a cost might be partly attributable, attribute part of it.
A conservative number that holds up is far more useful than a large one that gets picked apart, because the argument does not survive the first challenged assumption. And in most cases the conservative number is still large enough to change the conversation.
The comparison that actually lands
Once you have a figure for one departure, two comparisons make it useful.
Against the annual total. One departure multiplied by the number of people who left last year. Boards experience turnover as a sequence of individual events; the annual figure is what makes it a pattern.
Against the intervention. What would it have cost to keep them? Sometimes the answer is nothing you could have afforded, and that is worth knowing too. Often it is a smaller number than the departure cost, and the comparison is the argument — made in the currency the budget is written in.
That comparison is also where you find out whether the real driver is pay at all. If people are leaving over caseload, supervision, or a manager, a pay increase is an expensive fix for the wrong problem, and the exit conversations you have already had will usually tell you which it is — if anyone wrote them down.
What to do this month
Pick the last person who left from a role you have replaced more than once. Build the four buckets for that one departure, with every assumption written beside its number.
Take it to your next finance or board meeting as a single page. Not as a proposal — as a figure the organization did not previously have. The proposal comes afterwards, and it lands differently once the cost of the alternative is on the table.
The Turnover Cost Calculator — the four buckets, in your browser, with nothing you enter leaving your device — is free at the Turnover Cost Calculator. No registration, no email address.