Recruitment advertising is the only significant media budget in most companies that nobody treats as media. It gets approved as a hiring cost, spent by people who were never trained to buy attention, and reviewed by nobody who would recognise a bad buy. That is why it leaks.
The four things you can actually buy
Every recruitment advertising invoice is one of four purchases, and the difference between them is who carries the risk when the ad does not work.
| What you buy | You pay for | Risk sits with |
|---|---|---|
| A job board slot | Time on a page, usually 30 days | You |
| Pay per click | Someone arriving at the posting | Shared |
| Pay per applicant | A completed application | The board |
| Programmatic | Placement bought automatically across many destinations | Shared, and it moves |
A slot is the oldest model and the least honest one. You pay the same whether the role is a magnet or a ghost town, and the board has no reason to work harder on your behalf once the invoice clears. It is the recruitment equivalent of buying a billboard and never driving past it.
Who each pricing model favours, and it is not always you
Pay per applicant sounds like the safe choice, and for a hard role it often is. But read the definition of "applicant" in the contract before you celebrate. On some platforms an applicant is a person who clicked apply and abandoned the form on the second screen. You are then paying a fixed price for an incomplete record, and the platform has every incentive to make that first click as easy as possible.
Pay per click is the opposite trade. You carry the risk that the traffic is poor, but you also get the diagnostic value: if a thousand people click and four apply, that is not a traffic problem, it is a job description problem or an apply flow problem. A slot buy would have hidden that from you entirely.
The question to ask any vendor: what exactly triggers the charge, and can I see the raw count of people who reached the apply screen and did not finish? If they cannot answer the second half, you are buying blind.
Programmatic sits on top of these rather than beside them, spreading one vacancy across many destinations and moving spend towards whatever is producing. We covered the mechanics of that in detail in programmatic job advertising, including what it does and does not fix.
Cost per hire is the only number that settles an argument
Most recruitment advertising is reported on cost per applicant, because that is the number the platform can produce on its own. It is also the number most likely to mislead you, for the same reason cost per lead misleads advertisers who never check what closed.
Two campaigns, same spend. One delivers sixty applicants at a low unit cost and produces no hire. The other delivers twelve at triple the unit cost and fills the role in a fortnight. On the platform dashboard the first one wins by a wide margin. In the business, only the second one happened.
The fix is unglamorous. Someone has to carry the hire back from the applicant tracking system to the source that produced it, and keep doing that every month. Until that link exists, every optimisation decision is being made on the cheaper of two numbers rather than the truer one.
Three leaks that survive every budget review
Paying twice for the same person. Run a job board buy and a search campaign at once and a share of your applicants will arrive through both. The board bills you, and so does the click. Nobody deduplicates because the two invoices live in different systems and often on different budgets.
The apply flow. A candidate who clicks your ad and then meets a fifteen minute form, an account creation step and a request to re upload a CV they already attached is a candidate you paid for and lost. This is the single biggest leak in most accounts and it costs nothing to test. Apply for your own job on a phone and count the screens.
Postings that outlive the vacancy. Roles get filled and the advertising keeps running, because the person who filled the role is not the person who manages the spend. In a slot buy this wastes the remainder. In a programmatic or pay per click setup it keeps drawing money for weeks.
The tracking gap nobody owns
In a normal marketing account, the ad platform and the outcome live close together, and connecting them is standard practice. Recruitment breaks that by design. The ad platform sits with marketing or with an agency. The applicant tracking system sits with HR. The hire is recorded by a hiring manager who never sees either.
The practical consequence is that the ad platform is being asked to optimise towards the only signal it can see, which is the application. It will get very good at producing applications. Whether those applications become people who start on a Monday is a question it was never given the data to answer.
Closing that gap does not require new software in most cases. It requires the source to be captured on the application record and reported back on a fixed schedule, the same way an offline conversion import works in any lead generation account.
Setting a number you can defend
Two anchors make this conversation shorter. The first is what the vacancy costs while it stays open: lost output, overtime, the work not being done, the customer not being served. Most finance teams can produce that figure in an afternoon and it is usually larger than anyone expects.
The second is what you currently pay an external recruiter for the same role. If a contingency fee runs to a meaningful share of first year salary, then the advertising budget that would avoid one such fee is not a cost centre, it is the cheaper of two options.
Set the budget against those two numbers rather than against last year's advertising line, and the question stops being how little you can spend and becomes how quickly you can fill the seat.
Common questions
The questions employers ask most often about this.
Is recruitment advertising different from recruitment marketing?
They overlap and get used loosely. Recruitment advertising is the paid part: buying placement for a specific vacancy. Recruitment marketing is broader and includes the employer brand work that makes the paid part cheaper over time. If you need someone to start next month, you are buying advertising.
Should a small company use job boards or run its own ads?
Start with boards for roles where candidates are actively looking, because that is where the intent already exists. Run your own search and social campaigns when the person you want is employed and not looking, which is most senior and specialist hiring. The two answer different problems and the mistake is expecting either to do both.
Why did our applications go up but our hires did not?
Almost always because the campaign is optimising towards applications and nothing tells it what a good one looks like. Feed hires back to the source, even manually in a spreadsheet, and the pattern usually becomes obvious within one hiring cycle.
How long should a job ad run before we judge it?
Long enough to collect a meaningful number of clicks rather than a meaningful number of days. A role with heavy search demand can tell you something within a week. A niche role may need a month before the numbers mean anything at all.
