Cost-per-hire is easy to measure, so it gets measured. Cost-of-vacancy is harder, so it mostly does not — and it is generally the bigger figure by a wide margin.
A rough calculation you can do today
Take the role’s fully loaded annual cost — salary plus benefits and overhead, call it 1.25x salary. Divide by 52. That is the labour you are not getting per week.
For a $120,000 engineer, that is roughly $2,880 a week.
But labour foregone is the small half. The larger half is output foregone, which depends on what the role does:
- Revenue-adjacent roles (sales, delivery, billable consulting) — model the quota or utilisation directly. A billable consultant at $150/hour on a 70% utilisation target is roughly $4,200 a week of unbilled capacity.
- Throughput roles (engineering, operations) — the cost lands as delay on everything the team ships, plus overtime and error rates from colleagues absorbing the work.
- Control roles (compliance, finance, QA) — the cost is risk, which is invisible right up until it is not.
The compounding part
Vacancies are not linear. Weeks one to four, the team absorbs the work and looks fine. Weeks five to twelve, the absorption itself becomes the problem: your strongest people are covering, their own work slips, and the good ones start updating their own resumes. A long vacancy in one seat is a reliable predictor of a second vacancy on the same team.
What this changes about hiring decisions
Once you have the weekly number, several trade-offs stop being arguments:
Interview loop length. A five-stage loop spread over five weeks is not “thorough” if the seat is costing you $6,000 a week. Compress the calendar, not the rigour — same stages, same week.
Candidate quality bar versus availability. Holding out four extra weeks for a marginally stronger candidate is a real spend. Sometimes worth it. Now you can say how much.
Agency fees. A placement fee looks like a large one-off next to zero. Next to eight weeks of vacancy it frequently is not. That is not an argument for using an agency on every role — it is an argument for knowing which number you are comparing against.
One caveat
Do not use this to justify hiring the wrong person quickly. A mis-hire costs more than the vacancy did, and unwinding it takes months. The point is to stop treating delay as the safe option, because it is simply a cost you are not writing down.
