Cost per filled shift: the number most staffing firms have never calculated
Gross margin per placement is a reporting number. Cost per filled shift is an operating number — and it is the only one that tells you which workflow to automate first.
Most staffing operators know their gross margin per shift to the cent. Far fewer know what it costs them to produce that shift: the recruiter minutes, the coordinator touches, the confirmation calls, the timesheet chasing, the payroll corrections, the client email at 6am. That cost sits inside overhead, spread across salaries, and never appears next to the shift it belongs to.
It matters because overhead is where high-volume staffing margin actually dies. Job&Talent reported managing 63 workers per internal FTE against an incumbent average closer to 19 (source). That is not a sales difference or a rate-card difference. It is a cost-to-serve difference, and it compounds with every worker added.
The decomposition
Cost per filled shift is fully loaded internal cost divided by shifts filled, then broken into the activities that create it. The breakdown is what makes it actionable:
| Cost line | What you measure | Where it hides |
|---|---|---|
| Sourcing and screening | Recruiter minutes per qualified candidate, including unanswered calls and repeat dials | Recruiter salaries; job-board spend re-buying candidates already in your database |
| Requirement intake | Coordinator touches per client order, from first message to structured requirement | WhatsApp groups, inbox triage, spreadsheet retyping |
| Confirmation and no-show recovery | Chase cycles per shift; backfill cost when a worker does not arrive | Evening and early-morning coordinator time; client credits |
| Attendance to payroll | Corrections per 1,000 timesheet lines; days from shift end to invoice | Payroll rework, duplicate payments, unbilled hours |
| Worker support | Contacts per worker per month and average handling time | Support headcount; churn caused by unanswered pay questions |
A worked example
Take an operator filling 20,000 shifts a month with 60 internal staff at a fully loaded $3,000/month each. Total internal cost is $180,000, so cost per filled shift is $9.00. Time-allocation sampling across those 60 people typically lands somewhere near this shape:
| Activity | Share of internal time | Cost per filled shift |
|---|---|---|
| Screening and re-screening | 31% | $2.79 |
| Requirement intake and client updates | 22% | $1.98 |
| Confirmations, no-show recovery, backfill | 18% | $1.62 |
| Timesheet chasing, payroll corrections, invoicing | 17% | $1.53 |
| Worker support and everything else | 12% | $1.08 |
The shares are illustrative until measured in your business — the point is what the arithmetic does to the conversation. A 60% reduction in screening effort is worth about $1.67 per shift, or $400,000 a year at this volume. That is a number a CFO can approve, and it is a completely different conversation from "AI will make recruiters more productive."
Leakage belongs in the same number. Payroll leakage — duplicate payments, rounding drift, unbilled overtime, workers paid for shifts nobody can evidence — is cost to serve even though it never appears on a payslip line called "overhead." On one 300,000-worker deployment, closing the attendance-to-payroll loop cut leakage by roughly 90%.
How to measure it in two weeks
- Fix the denominator. Shifts filled, not placements, not orders received. One worker attending one shift is one unit.
- Sample time, don't survey opinion. Two weeks of activity sampling across recruiters, coordinators and payroll staff, at the task level. Ask what happened yesterday, not what usually happens.
- Count touches from system logs. Call records, message threads, ATS event history and timesheet edit trails give you touch counts per order and per shift without asking anyone to fill in a form.
- Price the exceptions separately. No-show backfills, payroll corrections and credit notes are not average cases; they are the tail that pays for automation.
- Rank by saving, risk and payback. Every candidate workflow gets an annual saving, a failure mode, and a number of weeks to payback. Automate the top of that list, not the most interesting problem.
Measure before you build. If a workflow cannot be measured before and after, it cannot be proven — and unprovable automation is where staffing technology budgets go to die.
Questions operators ask
What data do I need before measuring cost per filled shift?
Payroll cost for internal staff, shifts filled for the same period, and read access to call, message, ATS and timesheet logs. Two weeks of activity sampling covers the rest. You do not need clean data — you need enough to establish the shape of the cost.
How is this different from cost per hire?
Cost per hire measures the acquisition of a person once. Cost per filled shift measures the recurring cost of putting that person to work, which is where high-volume and gig staffing margin is actually consumed. A firm can have a good cost per hire and still lose money on every shift it serves.
Does the number include recruiter commission and job-board spend?
Include job-board and sourcing spend, since re-buying candidates you already own is a real cost line. Treat variable commission separately if it is genuinely tied to revenue rather than to effort — otherwise it distorts the automation ranking.
What does a good cost per filled shift look like?
There is no universal benchmark; segment, geography and pay frequency move it too much. The useful comparison is against your own baseline and against workers managed per internal FTE, where the published spread between AI-led operators and incumbents is roughly three to one.
We will measure it with you in two weeks.
The Ops Audit is $7,500 fixed and free if it does not identify ten times its fee in annual savings. You leave with the decomposition, the ranked list, and the two workflows worth fixing first.
Request a free 30-minute teardown →