What it actually costs to replace someone
Recruitment fees are the visible part and rarely the largest. The rest is absorbed invisibly, which is why retention loses budget arguments.
Retention measures compete for budget against things with clear returns, and they usually lose, because the cost they avoid is never quantified. The replacement cost appears in no report: it is distributed across a dozen people's weeks and never totalled. For broader independent guidance on this point, consult Gallup Workplace.
Costing it once, roughly, changes the conversation. The number is invariably larger than expected.
The four components
- Direct recruitment — advertising, agency fees if used, background checks, and the management time consumed by shortlisting and interviewing. This is the only part that appears in accounts, and in most small companies it is a minority of the total.
- Vacancy cost — the work not done, or done by colleagues at the expense of their own. It runs from the resignation to the replacement's start, typically two to four months, and it is where overtime and quiet quality decline accumulate.
- Ramp-up — the gap between the new person's cost and their output. Even a strong appointment is well below full productivity for months, and the trainer is below theirs while training.
- Knowledge loss — relationships, undocumented process, the reasons behind past decisions. Hardest to quantify and often the most expensive, particularly for long-tenured people.
Add the four components at your own numbers before reaching for a published multiple. For a skilled role in a small company the total commonly lands somewhere between a third and a full year's salary — but the point is to produce your own figure, because it is yours that will be believed internally.
Cost it for a real, recent departure
Abstract figures do not persuade anyone. Take someone who actually left in the last year and work through it: how long the vacancy ran, who covered, what was dropped, when the replacement reached full productivity, what had to be reconstructed. For a related product-level perspective on productivity and efficiency, see this guide.
The exercise takes an afternoon. Its output is a specific number attached to a specific person, which is considerably harder to dismiss than an industry average.
Some departures are more expensive than others
Replacement cost is not uniform, and knowing where it concentrates tells you where retention effort is worth spending.
The expensive departures are the sole holder of a critical skill, anyone with the primary customer relationship, long-tenured people carrying undocumented knowledge, and anyone whose exit is likely to be followed by others. The last is worth watching: departures cluster, and the first one in a cluster is often the one that made the others thinkable.
Compare against what prevention costs
Once there is a number, the comparison becomes concrete. Against a replacement cost of a third of a salary, the alternatives are cheap: an above-market adjustment for someone underpaid relative to recent hires, the cross-training that removes the single-point-of-failure burden, the equipment that has been promised twice, a fortnightly half hour of the manager's time.
None of these are guaranteed to retain anyone. All of them are inexpensive relative to the thing they might prevent, and that ratio is the argument.
Not every departure is worth preventing
This is worth saying plainly. Some turnover is healthy, some appointments should end, and some people should move on for their own reasons and with the employer's support.
The costing exercise is not an argument for retaining everyone. It is an argument for knowing what an avoidable departure costs, so that the decision to spend nothing on preventing it is a decision rather than an oversight.