Most employers can tell you what they paid an agency last year. Very few can tell you what hiring actually costs them, because the visible fee is frequently the smallest of four numbers.
The four are: the direct cost of filling the role, the internal time consumed, the cost of the role being empty, and the cost of hiring the wrong person. The last two are almost always larger than the first two, and they are the two nobody counts.
1. Direct cost
The money that leaves the business to fill the role.
Agency fees, traditionally quoted as a percentage of first-year salary, commonly in the range of fifteen to twenty-five percent and higher for search. On a mid-level salary that is a substantial sum for a role that may have had reasonable supply.
Fixed-fee alternatives price the work rather than the salary, which decouples the cost from what you happen to pay the person. HRM8 works this way: Shortlist from US$990 and Recruit from US$3,490.
Advertising. Premium job board placements, which vary widely by market and board.
Assessment. Priced per candidate, so the number depends on how many you put through. HRM8's Assessment Hub runs from US$25 per candidate.
Applicant tracking. Either per job, from US$195, or a subscription from US$295 a month. A free tier covers a single open role.
Background and reference checking, where applicable.
2. Internal time
The cost everyone knows exists and nobody calculates.
A typical mid-level hire consumes: several hours writing the job description and getting it agreed, several more screening, a screening call per shortlisted candidate, then interviews involving two to four people for an hour each across two or three rounds, plus debriefs, scheduling, reference checks and the offer conversation.
Multiply that by the loaded cost of the people involved. For a role involving a hiring manager and two interviewers on professional salaries, the internal time frequently exceeds the advertising cost several times over and can approach a fixed recruitment fee on its own.
This matters for the decision in a specific way: using external support does not eliminate internal time, it reduces it. The screening and shortlisting hours largely go. The interviewing and decision hours do not, and should not.
3. The cost of the role being empty
The largest number in most cases, and the one least often on the spreadsheet.
Work not being done. Colleagues covering, doing their own jobs less well. Projects delayed. For revenue-generating roles, revenue not earned. For customer-facing roles, service degrading in ways that produce churn later.
A crude but workable estimate: the role's fully loaded monthly cost is a floor for its monthly value, or the business would not be hiring. For revenue roles, use the actual contribution.
This reframes the whole decision. If a role is worth a meaningful sum per month and the choice is between filling it in six weeks with help or four months without, the fee is small against the difference. It also explains why slow internal processes are expensive even when they look free: every week of delay has a price, it is simply not invoiced.
4. The cost of the wrong hire
The number that dominates everything else, and the hardest to estimate precisely.
It includes the original hiring cost, repeated. The salary paid during a period of low productivity. The management time spent trying to make it work, which is considerable and usually spread across months. The disruption to the team, including the risk of losing good people who are carrying the load or watching a problem go unaddressed. Any exit process. And then the whole cost of hiring again, now under more pressure.
Widely cited figures put this in the range of several months' to more than a year's salary, and the estimates vary enough that quoting a single number is unwise. The conclusion does not depend on the precision: it is a large multiple of anything you would spend to improve your odds.
Which is the whole argument for structured hiring, assessment, and taking the time to get the decision right. Not because rigour is virtuous, but because the downside is asymmetric.
Making the comparison properly
For a given role, estimate all four.
Running it in-house: advertising, assessment, ATS, internal hours at loaded cost, and the vacancy cost for your realistic time to fill.
Using support: the fee, reduced internal hours, and the vacancy cost for the shorter expected time to fill.
Then apply the honest question: does one approach materially change the probability of a good hire? For a role with healthy supply, probably not, and you should run it yourself. For a scarce role where the right person is not applying to adverts, it changes it substantially.
The metric to stop using
Cost per hire, averaged across an organisation, is close to useless as a management number.
It can be improved by doing worse hiring: advertise cheaply, skip assessment, interview once, decide fast. All of that reduces cost per hire and increases the cost of wrong hires, which does not appear in the metric.
Better measures: time in each stage, which shows where delay actually accumulates; offer acceptance rate, which tells you whether your pay, proposition and speed are competitive; and twelve-month retention of new hires, which is the only measure of whether any of it worked.
A process that is slightly more expensive per hire and produces people who stay is not a worse process. Cost per hire cannot see that.
Common questions
What is the average cost per hire?
Published averages exist and are not useful for your decision, because they vary enormously by role, seniority, market and method. Estimate your own four components for the specific role in front of you.
How much do recruitment agencies charge?
Traditionally a percentage of first-year salary, commonly fifteen to twenty-five percent, and higher for search. Fixed-fee models price the work instead: HRM8's Shortlist starts at US$990 and Recruit at US$3,490.
What does a bad hire cost?
Estimates range from several months' salary to well over a year's, and the variance is wide enough that a single figure is not credible. It reliably exceeds any reasonable spend on getting the decision right.
How do you calculate the cost of an unfilled role?
Use the role's fully loaded monthly cost as a floor for its monthly value, or actual contribution for revenue-generating roles. Multiply by expected months open.
Count four numbers rather than one: direct cost, internal time, the cost of the role being empty and the cost of hiring the wrong person. The last two usually dominate, which is why slow processes are expensive even when they look free, and why rigour is cheap against the downside.