Headcount variance: separate staffing, timing and pay

Headcount variance is actual staffing minus planned staffing for the same date and business scope. Count employees and full-time equivalents separately, then use period-average FTEs to explain the cost of vacancies and late starts. A closing roster alone cannot explain a month's salary expense.

Define the count before comparing it

An employee count answers how many people are on the included roster. An FTE count answers how much scheduled working capacity those people represent. Two people working half of the company's full-time schedule count as two employees and one FTE.

The Census Bureau's employment glossary illustrates the underlying hours-based principle: it converts part-time hours into a full-time equivalent using the applicable standard hours. Its survey covers governments; a company's management report must state its own schedule and population rather than imply that a survey definition governs its internal budget.

Keep a snapshot measure and a period measure:

• Closing headcount and closing FTE show staffing at the reporting date.
• Period-average FTE captures the portion of the period each included position was staffed.
• Expense uses that period exposure and the corresponding compensation basis.

For internal planning, a half-time employee present for half the month contributes 0.25 period-average FTE: 0.50 schedule FTE multiplied by 0.50 of the month. Calculate that from effective dates, using the same proration convention in plan and actual.

The formulas

Headcount variance = Actual closing employees − Planned closing employees

Closing FTE variance = Actual closing FTE − Planned closing FTE

Period FTE variance = Actual period-average FTE − Planned period-average FTE

Staffing expense effect at plan rate = Period FTE variance × Planned expense per FTE

The first three results are staffing measures, not dollar amounts. The final formula translates staffing exposure into an expense comparison while holding compensation constant. Use a separate rate for each comparable role or department; a company average can hide a shift between expensive and less expensive positions.

The broader variance analysis method explains the actual-minus-plan convention. Here a negative staffing result means below plan. Calling it favorable requires a separate judgment about cost, capacity and service.

Start with the staffing and expense exports

Use an approved position plan, an actual roster with effective dates, and the matching departmental expense summary. Join them by position identifier and department, not employee name alone. Preserve plan version, reporting date, employment category, scheduled FTE, start date and end date.

Include vacant planned positions in the plan file. Otherwise, matching only employees who appear in both exports deletes the very vacancies being measured. Map transfers once so the same employee does not appear in two actual departments on the closing date.

This September example is a US business. The roster and plan produce these measures; FTEs are shown to one decimal place, with no rounding needed for the calculations.

Closing people are one below plan: 23 − 24 = −1. Closing FTE is also one below, but average FTE is two below: 21.0 − 23.0 = −2.0. Engineering reaches its closing target even though a hire started halfway through the month. Customer service has both a vacancy and a half-month delay on another hire.

Translate staffing into expense without blaming pay

The simplified expense definition below includes base salary only. It excludes benefits, payroll taxes, bonuses and overtime. Within each department the supplied monthly expense per average FTE is comparable between plan and actual.

For this example, split staffing into closing capacity and timing. Closing capacity uses closing FTE variance at the planned rate. Timing is the remaining average-FTE difference at that same rate. The compensation effect uses actual average FTE multiplied by the actual-minus-plan rate difference.

This decomposition is an endpoint-and-period comparison, not proof of individual hiring events. The effective-date rows substantiate the timing explanation. For the wider expense reconciliation, including one-time charges and accruals, use analyzing headcount costs.

The checks that prove the result

First, every actual employee must belong to one closing department, and departmental counts must sum to the roster total. Next, average FTE must rebuild from the underlying effective-date fractions. Finally, the expense effects must reconcile: −6,000 − 8,000 + 1,900 = −$12,100, equal to $163,900 − $176,000.

Reconcile the actual salary total to the same expense accounts in the ledger. A cash payroll payment and an accrued salary expense can cover different dates. Finance owns that reconciliation; HR owns roster status and effective dates. Department managers confirm whether the gaps reflect approved vacancies, delayed hiring or changed capacity needs. FP&A's role connects those inputs to the wider operating plan.

Where it goes wrong

• Month-end people used as monthly exposure. Engineering would appear fully staffed all month, hiding its timing effect.
• Different populations. Counting contractors in actual while excluding them from plan produces an artificial hiring variance.
• A legal definition copied into planning. The IRS full-time employee definition applies specifically to employer shared responsibility provisions. It is not automatically the company's internal FTE denominator.
• An apparent saving treated as success. Vacancies can defer expense while increasing backlog. A favorable variance label needs context about the business result.

Compare your staffing files in Covirage

Bring your position plan, dated roster and matching departmental expense export. Covirage maps the columns, and its tools compute and check the staffing and expense differences; the external AI model explains the results with every figure traced to its rows. Create a free Covirage account and bring a sample of your own data.

See cost variance analysis for the supporting finance workflow.