How to Measure Recruitment Marketing: What to Instrument Before the First Dollar, Three Cadences, One Rule, and What You Can Prove to Finance

Last updated August 27, 2026 · The measurement system · The metrics themselves are on the recruiting analytics page; conversion on its own page; source tracking on the attribution page

Measuring recruitment marketing is a system, not a dashboard: instrument three things before the first dollar (an in-platform form, a source on every application, hires tagged in the ATS), read three numbers weekly per campaign (frequency, apply rate, cost per applicant with the fee inside), and roll them up monthly by campaign and quarterly by channel. One rule holds it together — report by campaign, never in aggregate — because in Boostpoint's 2026 Social Job Advertising Benchmark the top 10% of campaigns produced 57% of all applicants on 22% of the budget, and an average hides exactly that. What you can prove to finance is cost per applicant and where the money concentrated; what you can't prove from ad data is a cost per hire, and this page says why. Below: the instruments, the cadences, the rule, the proof, and two tools.

Source: Boostpoint 2026 Social Job Advertising Benchmark — 891 Boostpoint-managed campaigns on Meta (1,334 campaign-months), 2026; costs are what advertisers paid, inclusive of campaign management, and cover advertising only.

Book a Demo

A note on who's writing this

Boostpoint runs recruitment campaigns and reports them weekly, so the system on this page is the one we use, and a page recommending our own reporting practice is a page recommending us. The figures behind it are the published 2026 benchmark — 891 campaigns and 1,334 campaign-months, management fee inside — with the expensive end shown; the frequency threshold and the report-by-campaign rule are what the data drew, not what we prefer. We don't track hires, so nothing here is a cost per hire, and the section on what you can prove says plainly where our data stops and your ATS starts. The measurement plan builder is our judgment; the concentration calculator is arithmetic on your own campaigns.

The case for measuring less, made first

The honest objection to a measurement system is that most of what gets measured in recruiting changes nothing. A dashboard with fourteen tiles — impressions, reach, engagement, click-through, cost per click, cost per thousand, video views, saves — is a report of what happened, and none of those tiles tells a recruiter what to do on Monday. Time to fill goes up when the pool is thin and down when it's not, and a manager can't act on it. Source of hire is only as good as the source field, which is usually blank. So the case for measuring less is right: three numbers a week per campaign, read against a threshold, decide everything that can be decided. The case for measuring at all is the concentration finding — that the money was in a few campaigns and nobody could see which ones from the average — and that finding is only visible to an employer who instrumented the three things below before the campaigns ran.

What to instrument before the first dollar

Three instruments, in the order they pay, and none of them is a tool you buy. The application inside the platform. When the form lives in the ad, the platform's applicant count is real and the click-to-application rate — the number that explained 70% of cost variation in the benchmark — is visible. When the click leaves for a career site, the platform goes blind at the click and every later number is a guess; the reference case, 60,889 clicks and zero tracked applications, is on the attribution page. A source on every application. The integration writes it, a tagged URL carries it, the ATS field requires it; the vocabulary is one list, and "Facebook," "FB" and "social" are one source, not three. Hires tagged in the ATS. Not for the platform — it will never see them — but for you, because the only cost per hire that exists is applications with a source, divided into hires with the same source, in your own system. Instrument those three and the rest of this page is arithmetic; skip any one and there's a layer you're reporting on faith.

Three cadences, and what each one decides

Measurement has a rhythm, and the rhythm is set by how fast each number can change. Weekly, per campaign — three numbers, one threshold. Frequency, because it crosses the cliff in a fortnight: campaigns under 1.5 impressions per person ran a median $7.85 per applicant in the benchmark, campaigns over 4.0 ran $26.78, click-through fell from 1.69% to 0.94% across the range, and 33% of all budget ran above 3.0; the operating rule is to act at 2.5 — rotate the creative, widen the audience, or cap the budget. Apply rate, because it tells you whether the form or the first line is leaking (under 10% costs $23.13 to $53.77 per applicant; over 20% costs $4.41 or less). And cost per applicant with the management fee inside, because it's the number the other two explain. Nothing else weekly. Monthly, by campaign — the concentration cut. Sort every campaign by cost per applicant, read the top and bottom thirds, and ask the funnel question of each expensive one: is it frequency, or is it the form? This is the read that finds the 57%-on-22%, and it only works campaign by campaign. Quarterly, by channel — the reconciliation. The platform's count against the ATS's tagged count (coverage), the unknown-source share, hires and 90-day survivors by source, and the reconciled cost per applicant by channel with the fee inside — which is what the next quarter's budget is built on. The quarterly sheet is described on the attribution page; the budget it feeds is on the budget page.

The one rule: report by campaign, never in aggregate

If this page had to be one sentence, it would be this one. Across 1,334 campaign-months, the top 10% of campaigns produced 57% of all applicants on 22% of the budget, and the top 30% produced 83% on 53%; the bottom of the distribution, read the other way, is a large share of spend producing very little. An employer looking at the account-level average — spend divided by applicants across everything — sees a number near the middle and learns nothing, because the middle is where almost no campaign actually sits. Report by campaign and the same data says: these two campaigns are doing the work, these three are burning money at a frequency of 4, and this one has a portal behind it. The same rule applies to every roll-up. By role, because a caregiver campaign at 31% apply rate and a nurse campaign at 11% average to a number that describes neither; by market, because a metro and a small town on the same template have different circles (the frequency trap); by channel, because a board's clicks and a social campaign's applicants are different funnels. The calculator below does the concentration cut on your own campaigns; enter eight and read where the money went.

Recruitment marketing metrics: the six that decide, and the ones that don't

The metric list on the analytics page is six long — cost per applicant with the fee inside, apply rate, frequency, reach, cost per campaign, and the concentration cut — and the discipline is in what's not on it. Impressions and reach describe how many people the auction served; useful for one thing, computing frequency, and not a result. Click-through rate matters (it explained 30% of cost variation) but it's a symptom, not a lever; the first line fixes it. Cost per click is a board's metric, and a click is not an applicant — the translation at whatever apply rate you observe is on the job board ROI page. Engagement, saves, shares and video views are not recruiting metrics at all. Time to fill is an ATS metric that measures the pool and the process together and can't be acted on by an advertiser. And cost per hire, the metric everyone asks for, is the one no ad platform can supply, because the hire happens where the platform isn't looking. A measurement system that reports the six and refuses the rest is not less rigorous; it's the only version a recruiter will still be running in six months.

Proving recruiting ROI: what you can show finance, and what you can't

Finance wants one number and there isn't one, so the honest version is a short list of what's provable and a shorter list of what isn't. Provable from ad data with the fee inside: cost per applicant by campaign, by role and by channel; how many applicants a dollar buys; where the money concentrated and what was cut; the apply-rate band each campaign sits in and what moving it would have returned. Our benchmark's shape is the reference — a median $13.88 per applicant, $8.02 volume-weighted, a normal campaign on $334 a month for 20 applicants and 6,061 people reached — and yours is the same shape on your own numbers. Provable only from your ATS: hires by source, cost per hire by channel, 90-day survival by channel (why that column matters), and the return on a seat — which needs the value of a filled shift or a filled truck, a number finance already has. Not provable by anyone: a cost per hire from ad spend divided by an assumed hire ratio, which is what most "recruiting ROI" reports are. The shape of a finance conversation that works: here's what an applicant cost, here's where the money concentrated, here's what we cut, and here's the ATS's count of who we hired from each channel — with SHRM's "nearly $4,700" average cost per hire (April 2022) as the cross-industry context for the size of the seat, not as a number of ours. What the outsourced version of this reporting should hand you every week is on the services page.

When the measurement system is more than you need

Three cases. If you run one campaign for one role, read it weekly against the frequency threshold and the apply-rate band and skip the monthly and quarterly layers; there's nothing to concentrate. If you hire a handful a year, the ATS source field will say "referral" and the honest measurement is a conversation with the person who referred them. And if the ATS can't take a source at import, build the weekly and monthly layers anyway — they need only the platform — and treat the quarterly reconciliation as the reason to fix the intake, not as a spreadsheet to maintain by hand. What no employer should skip is the instrument that costs nothing: keep the form where the ad is.

Frequently asked questions

Boostpoint figures come from the 2026 Social Job Advertising Benchmark — 891 campaigns and 1,334 campaign-months, costs inclusive of campaign management: top 10% of campaigns 57% of applicants on 22% of budget, top 30% 83% on 53%; conversion explaining 70% of cost variation, click-through 30%, CPM 18%; 41% of budget in campaigns converting under 10% returning 11% of applicants; apply-rate bands under 5% $53.77, 5–10% $23.13, 10–20% $11.11, 20–35% $4.41, over 35% $1.61; frequency under 1.5 $7.85, over 4.0 $26.78, 33% of budget above 3.0, click-through 1.69% to 0.94%, act at 2.5; median $13.88 per applicant, $8.02 volume-weighted, percentiles $2.91 and $66.45; a normal campaign $334 a month, 20 applicants, 6,061 reached. SHRM, average cost per hire "nearly $4,700," April 11, 2022 — cross-industry context, not a Boostpoint figure. The recruiting-roi glossary entry redirects to the recruiting analytics page; nothing on this page duplicates it. Boostpoint is priced as a product subscription plus ad spend: the subscription is priced on job volume and quoted on a call, ad spend is separate and recommended at a minimum of $750 a month per job category, there is one bill, through Boostpoint, and every cost-per-applicant figure we publish includes campaign management. The plan builder is Boostpoint's practice; the concentration calculator computes on the reader's campaigns. Cost per applicant is not cost per hire; we do not track hires and publish no cost-per-hire figure. A sample of Boostpoint campaigns, not an industry-wide study.