Cost of Vacancy: How to Calculate What an Unfilled Role Costs You Per Day

An unfilled role costs you three things every day it stays open: the output nobody is producing, the money you spend covering for it, and the hours your current team loses absorbing the work. Add those three and divide by the working days in the period. That is your cost of vacancy per day, and for most roles it is a bigger number than anything you would spend to fill the role faster.

Here is how to run it, with two worked examples you can copy into a spreadsheet in about ten minutes.

The formula

Cost of vacancy per day = (lost output + coverage cost + team drag) / working days

Use 260 working days a year, or 21 a month. Be roughly right rather than precisely wrong. Nobody is auditing this. You are trying to turn "the role is still open" into a number your CFO can act on.

Lost output is what the role produces when someone sits in it. For a quota carrying role, it is the quota. For a role tied to a delivery date, it is the value of the thing that ships late. For a role that keeps something from breaking, it is the cost of it breaking.

Coverage cost is real money leaving the building right now: contractors, overtime, a temp, a manager working weekends. This one is easy, because it already shows up in the ledger.

Team drag is the piece most calculations skip, and it is usually the largest. When a role is open the work does not stop. It gets redistributed to people who already had a full week.

Worked example: a role that carries a number

Say you are hiring an account executive at $150,000 base carrying a $1.2 million annual quota. Those are your two inputs. Everything after that is division.

$1.2 million / 260 working days = about $4,615 of production per day the seat is empty.

Ninety days open is roughly $415,000 of quota that nobody carried. Then add your own ramp. If your reps take three months to reach full production, the person you hire in December is not at target until March, so a hiring decision you delay by 30 days moves your revenue by 30 days. The delay is not absorbed. It gets pushed to the end.

You will get an objection here: unfilled quota gets picked up by the rest of the team. Some of it sometimes does. Discount the figure if you want, take half of it, and the answer is still over $2,000 a day. The argument does not need the full number to hold.

Worked example: a role that does not carry a number

Most open roles are not sales. An engineer, an ops lead, a controller. There is no quota to divide, so people skip the exercise entirely and the role quietly stays open another quarter.

Do this instead. Take the loaded cost of the role, salary plus benefits and taxes, usually somewhere near 1.25 times base. On a $150,000 engineer that is about $187,500 a year, or roughly $720 a working day. That is not the cost of vacancy. That is money you are not spending, and it is exactly why finance sees an open role as an underspend.

The cost sits on the other side of the ledger:

• The work still happens, by other people. If four teammates each give up 20 percent of their week covering, you have diverted close to a full role's worth of senior capacity to backfill a gap. Price that at their loaded cost, not at the open role's.
• The thing that slips. Ask the hiring manager one question: what is not shipping because this role is empty, and what is that worth? They will answer immediately. They have been thinking about it for weeks.
• The person who leaves. Coverage has a shelf life. The second resignation caused by the first vacancy is the most expensive line in this whole calculation, because now you are running two searches instead of one.

You will not get a precise figure. You will get a defensible range, which is all you need, because the number currently in the room is zero.

The number finance is using is wrong

An open role shows up in the budget as an underspend. At quarter end it looks like a win. There is no line item anywhere in the P&L called "revenue we did not earn because the seat was empty for 100 days," so the cost stays invisible while the saving is visible and celebrated.

Your job is to make the invisible side visible before the next planning cycle rather than during it. Bring one page: cost of vacancy per day, days open to date, cumulative cost. Three numbers. That page moves approval conversations further than any argument about candidate experience ever will.

What to do once you have the number

Cost of vacancy per day is a budget for speed. It tells you how much you are allowed to spend to close the gap faster, and it usually reveals you have been arguing about the wrong line.

Take the sales example at $4,615 a day. A contingency search on that $150,000 role typically runs 20 to 25 percent of first year base, so $30,000 to $37,500. That fee is the equivalent of roughly seven or eight days of vacancy. If it fills the role three weeks sooner than you would have on your own, the fee has covered itself and the debate about the percentage is beside the point. We break down what that fee actually buys in our cost to hire through a recruiter guide.

The same logic prices everything else on the table: a sourcing subscription, a contractor bridging the gap, cutting a step out of your interview loop, paying more at offer. Each becomes a trade you can price rather than a preference you argue about.

Two cautions. First, speed you buy at the top of the funnel does nothing if the delay lives somewhere else. Most stalled roles are not short of candidates, they are short of feedback, and there is a specific way to find where the days are going in how to find the leak in a role that has been open 90 days. Second, one recruiter can only carry so many searches at once, and adding a tenth open role to their list slows the other nine. The hours math on recruiter capacity is worth running next to this one, because the two numbers tend to explain each other.

Put it on the weekly report

Cost of vacancy is not a one time exercise for budget season. Calculate it once per open role, then show cumulative cost next to days open in whatever your weekly hiring review already looks like.

Something shifts when a line reads "Senior Engineer, 74 days, $53,000" instead of "Senior Engineer, still open." Interview feedback that used to take five days arrives in one. The salary band question gets settled. The debrief gets scheduled. Nothing about the search itself improved. The cost just stopped being invisible.

If you run the numbers and find the constraint is capacity, that roles sit because nobody has the hours to work them, that is the problem AI Talent Partner is built for. It is a flat monthly subscription running continuous sourcing and outreach across three concurrent roles at the base tier, $3,000 a month plus a $3,000 setup, so the first year is $39,000 and the second is $36,000. That is roughly what a single contingency placement on a $150,000 role costs, except it covers three roles at once and does not charge you again on the fourth hire. You can see how it works at nextchaptertalent.ai. And if you would rather just run the arithmetic yourself first, do that. The number is the point.

FAQ

Is there a standard cost of vacancy benchmark I can use?

Not a useful one. Any figure quoted as an industry norm is hiding an enormous range, because a quota carrying sales role and a back office role at the same company can differ by an order of magnitude. Calculate it per role using your own quota, loaded cost and coverage numbers.

Should cost of vacancy include recruiting spend?

No. Keep them on opposite sides. Cost of vacancy is what staying open costs you. Recruiting spend is what closing it costs. Comparing the two is the entire value of the exercise, and folding them into one figure destroys it.

How do I calculate this for a role that has never been filled?

Use the business case that got the role approved. Someone justified it with an expected output, a revenue target, or a capacity gap, and that justification is your lost output input. If nobody can state it now, that is worth knowing too, because it means the role may not be the priority the calendar says it is.