The cost of a bad hire vs. an agency fee

Every staffing conversation eventually arrives at the same objection: the agency rate is higher than the wage. That is true, it is always true, and on its own it is not an argument. The comparison that matters is not markup versus wage. It is markup versus the fully loaded cost of getting the hire wrong — plus the cost of the seat sitting empty while you try again.

Most employers have never built that second number. This piece builds it, with the arithmetic exposed so you can substitute your own inputs.

1. Why the usual comparison is wrong

Set side by side, a direct hire at $22 an hour looks obviously cheaper than an agency bill rate on the same role. But those two numbers are not measuring the same thing.

The $22 excludes your employer contributions, your workers' compensation premium, your vacation and statutory holiday accrual, and every hour of your own staff's time spent finding and onboarding the person. It also assumes the hire works. The agency rate includes all the statutory items, includes the recruiting, and prices in the risk that it doesn't work.

So the honest comparison has three columns, not two: a direct hire that succeeds, a direct hire that fails, and an agency placement. And the only way to compare them is to know how often the second column happens to you. Most operations do not measure this, which is precisely why the objection feels so strong.

2. The seven costs of a hire that fails

When a general labour or warehouse hire washes out at week six, the cost is not "a few weeks' wages." It is seven separate things, and five of them never appear on an invoice.

1. Sourcing and screening time. Writing and posting the ad, sifting applications, phone screens, no-shows to interviews, interviews, reference calls. Charged at the real hourly cost of whoever did it — usually a supervisor or an office manager, not a recruiter.
2. Onboarding and paperwork. Setup, payroll enrolment, safety orientation, equipment issue, badge, boots.
3. Training and ramp-up. The weeks where the person is being paid full rate and producing at a fraction of it, plus the experienced worker whose output drops while they teach.
4. Supervisor attention. Correcting, re-explaining, checking work, and eventually documenting problems. This is the cost employers most consistently underestimate.
5. Errors, rework and scrap. Mis-picks, damaged product, a mislabelled pallet that comes back as a customer complaint. In food or pharma-adjacent work, one documented error can cost more than the whole hire.
6. Team drag. A struggling or disengaged worker on a line slows the line. Where the work is paced or paired, this is real and largely invisible.
7. The exit and the do-over. Termination administration, final pay, possible unemployment implications, and then the entire sourcing cycle again from zero — while the seat is empty.

3. A worked example

One warehouse general labour role. Nominal wage $22 an hour. The hire fails at the end of week six. All figures illustrative.

That figure excludes wages paid, which you would have paid anyway for work performed. It is pure loss.

And note which line is largest: the empty seat. Which is the part most cost-of-hire discussions leave out entirely.

4. The cost of the seat, per day

This one you can calculate properly, because it's operational rather than speculative. Pick whichever version fits your floor:

• Output-based. Units, lines, orders or pallets that position produces per shift, times your contribution margin per unit. The cleanest number if you have it.
• Overtime-based. If the gap is covered by existing staff on overtime, the cost is the premium: the extra half-rate on every hour covered, plus the fatigue that shows up as errors and absence two weeks later.
• Service-based. If the gap is not covered, the cost is late shipments, missed appointment windows, chargebacks and the customer relationship. Usually the biggest of the three, and the hardest to book.

In the illustrative example above, an unfilled warehouse position was costed at roughly $210 a day in lost throughput and overtime premium. Eleven working days of vacancy is $2,300 — and eleven days is optimistic for a direct hire, because posting, screening and notice periods rarely compress below two weeks.

Once you know your cost per day of vacancy, speed stops being a convenience and becomes a line item. That is the number that changes the agency conversation.

5. What the markup costs on the same role

Now the other side. Same role, same $22 pay rate, filled through an agency for a twelve-week stretch while you decide whether to convert.

So on these illustrative numbers, twelve weeks of agency cover costs somewhere near $1,700–2,100 more than running the same person on your own payroll — against roughly $6,700 if the direct hire fails, and against an empty seat that starts costing on day one instead of week three.

Put differently: the arrangement pays for itself if it prevents roughly one failed hire in three. Whether it does depends entirely on your current failure rate — which is the number you should go measure before believing anyone, including us.

6. Where this argument breaks down

Three cases where the arithmetic goes the other way, and we would tell you so:

You have a strong internal pipeline. Some employers have a referral culture and a waiting list. If you can fill a floor role in four days from people who already know the plant, your vacancy cost is small and your failure rate is low. Use it. Agency spend on top of that is waste.

The role is genuinely permanent and stable. A single, well-defined, year-round position with predictable volume should end up on your payroll. Agency is for uncertainty, speed and peaks — not for indefinitely renting a permanent job, which costs more over a year and gives the worker no reason to stay.

Your real problem is retention, not recruitment. If you are replacing the same position four times a year, an agency will fill it four times a year and you will pay four times. The cause is upstream — shift pattern, first-week experience, supervision, commute. Fix that first; see warehouse turnover. An honest agency will tell you this. It costs us a placement and saves you the money.

7. Cutting the failure rate, whichever route you take

Most of what determines whether a hire sticks costs nothing:

• Describe the job accurately, including the unpleasant parts. Cold room, real lift weights, concrete floors, 5 a.m. start. Filtering honestly at the top costs you a few applicants and saves you the whole loss column.
• Check the commute before the offer. Not the driving time — the actual transit route at the actual shift start time.
• Put someone specific in charge of the first morning. By name, and present when the person arrives.
• Set a 30-day checkpoint with written criteria. Decide at day 30, not at month four when you've absorbed three months of drag.
• Have a real conversation at day two. Almost every wash-out was visible on day two and nobody asked.

8. The one number to start tracking

If you take one thing from this: track your 90-day survival rate by position. Of the people who started in this role over the last year, what fraction were still there at day 90?

That single ratio tells you whether your problem is sourcing, onboarding or the job itself, and it converts every argument on this page from opinion into arithmetic. If it's above 85%, your process works and you should be sceptical of agencies. If it's under 60%, you are already paying far more than any markup — just on lines of the P&L that don't have "recruitment" written on them.

---

Related: how temporary staffing actually works covers what the markup is composed of and what to ask before signing. Agency or direct hire works through which route fits which situation.