The short version: a favorable payroll variance can be evidence that your hiring plan failed. Split it into the vacancy you expected and the hiring that ran late, and decide up front who keeps the savings. Headcount planning in Abacum.

The University of Colorado Boulder tells its departments to assume a 4% vacancy savings rate when they build next year’s budget. The guidance is careful about it. The rate isn’t required, it’s an average across the whole university, and each unit is told to build the plan that works for them.

But the number goes in the budget. Before the year starts, in writing, they have said out loud that some roles will not be filled. Plenty of city budgets do the same thing. Yorba Linda carries a line item called Vacancy Factor, described in one sentence: the calculation builds in salary savings that will be realized during the fiscal year for staffing vacancies.

Corporate FP&A almost never does this. We model compensation to the dollar. Every approved role starts in January and stays filled for twelve months. Then we get the savings anyway, and we book them as a favorable variance.

Same money. One version is a plan. The other is a surprise you take credit for.

1. Your headcount plan already has a vacancy assumption

You’ve already made the assumption. You made it by leaving the field blank. If you never picked a vacancy number, the plan assumes zero, and zero is the one value you know is wrong.

That gets expensive at year end, because a favorable payroll variance and a hiring plan that never happened look identical on a P&L. Both are money you didn’t spend on people.

Your CEO reads it as cost discipline. Sometimes that’s exactly what it is. Often it’s a handful of roles that took a quarter longer to fill than planned, in the function you least wanted to slow down. Without a number to compare against, you can’t tell the two apart. You are reading two figures that happened to net, not a result.

2. Separate the vacancy you expected from hiring that ran late

What changed things for me was splitting one number into three. This is the core of headcount planning as a discipline rather than a spreadsheet tab.

The hiring plan. Which roles you meant to fill, and when.

Expected vacancy. The lag you assume between approving a role and someone actually starting. If you never picked a number, you assumed zero.

Vacancy variance. How much longer the roles actually stayed open.

Most plans carry the hiring plan and nothing else. That’s why expected vacancy and vacancy variance end up being argued about in the spring, with no evidence in the room.

With all three you can say something worth saying in a board meeting. We’re $3M under on payroll, about a third of that was planned, and the rest is hiring running two months behind across the sales org.

The money is bigger than it sounds. Two and a half thousand people, call it $440M of fully loaded compensation, and a single point of vacancy is $4.4M.

3. Decide what happens to the savings

Here’s the failure mode, and I’ve watched it happen more than once. A VP has $400K sitting there from roles she hasn’t filled. She comes to you wanting contractors, or wants to pull next year’s software purchase forward. You look, the money is right there, and you approve it. Then the roles get filled in Q3 and the cost base comes back. Everybody behaved reasonably and you’re over budget.

You’ve got three options and all of them are defensible. Take the savings back centrally and redeploy them at the reforecast. Let the function keep whatever it didn’t spend. Or let them spend it, but make them come back through a reforecast to do it. What you can’t do is leave the question open, because then it gets answered for you by whoever asks first.

Tip. Pull requisition approval date against actual start date out of your HRIS, by department and role type, going back eight quarters. That gives you a vacancy assumption built from your own data. If you are building it outside a planning tool, our headcount planning template is a reasonable place to start. The variation between departments will tell you more than the average, and if it comes out dramatically under 4% I’d check the method before celebrating.

Vacancy savings will happen whether you model them or not. If you don’t model them, you can’t tell the difference between good cost control and a hiring plan that never happened.