Quick Answer
The recruitment metrics that matter to a staffing agency are different from the ones that matter to a corporate talent team. Agencies run on ratios — submission-to-interview, interview-to-placement, fill rate, redeployment rate and gross profit per desk. Time to fill and cost per hire matter too, but mostly as a story you tell the client, not as a way to run a desk.
Most recruitment metrics lists are written for in-house talent teams. They're fine lists. They just don't survive contact with an agency desk, where you have two customers, a margin to protect, and a client who thinks "time to hire" means whatever annoyed them last week.
This is the agency version: twelve metrics, each with its formula and benchmark source — plus an honest account of the benchmarks that don't exist at all.
There's also a cost-per-placement calculator further down. It shows its maths, and it only ever tells you about your own desk.
TL;DR — Key Takeaways
| Metric | Formula | Public benchmark? |
|---|---|---|
| Time to fill | Job order open → offer accepted (calendar days) | Yes — 39 days median, nonexec (SHRM 2026) |
| Time to hire | First contact → offer accepted | Partial |
| Cost per hire | (External + internal costs) ÷ hires | Yes — $5,475 nonexec (SHRM 2025) |
| Fill rate | Job orders filled ÷ job orders taken | No industry average |
| Submission-to-interview | Interviews ÷ submissions | None published |
| Redeployment rate | Contractors redeployed ÷ contracts ended | None published |
| Gross profit per recruiter | GP ÷ FTE recruiters | Locked in ASA member survey |
| Client NPS | Standard NPS | Yes — 45 industry (ClearlyRated 2024) |
Why Agency Recruitment Metrics Are Different
A corporate talent team optimises cost per hire. Every placement is a cost centre, and the job is to make it cheaper and faster. An agency optimises something else entirely: gross profit per recruiter per month. A placement isn't a cost. It's the product.
That changes which numbers deserve attention.
The two-sided funnel
On one side, a client pipeline: prospects, meetings, job orders, repeat orders. On the other, a candidate pipeline: sourced, screened, submitted, interviewed, placed. Either one can be the constraint. A desk with 40 open job orders and a thin candidate bench has a sourcing problem. A desk with a deep bench and three job orders has a business development problem. Same placement count, opposite fix. Corporate metrics only see the candidate side, because in-house teams never have to win the job.
Client conversations vs. management conversations
Here's a useful split. Time to hire is a client conversation — it's what they feel and what goes in the quarterly review. Submission-to-interview ratio is a management conversation — it's what tells you which recruiter needs coaching and which client keeps moving the goalposts. Mix them up and you end up coaching recruiters on a number that mostly depends on how fast the client books interviews.
The 12 Recruitment Metrics Staffing Agencies Should Track
Grouped into speed, conversion, money and retention. Where there's no benchmark, the metric says so — right there, not in a footnote.
Benchmark figures last checked October 14, 2026. SHRM, ClearlyRated and labour-market figures are reviewed annually.
Speed
1. Time to fill
Offer accepted date − job order opened date (calendar days)
SHRM defines it as "the number of days from when the job requisition was opened until the offer was accepted by the candidate" — calendar days, weekends and holidays included. For an agency the clock usually starts when the client gives you the job order, not when they opened it internally. Say which one you're using.
Benchmark: Median 39 calendar days for nonexecutive roles, across 4,600+ organisations (SHRM 2026 Recruiting Executives Benchmarking). That's an employer number: it includes everything your client does before and after you.
2. Time to hire
Offer accepted date − date the placed candidate entered the process
Time to fill measures the job. Time to hire measures the candidate. Clients conflate them constantly, which is how an agency that surfaced the right person in four days ends up blamed for a seven-week process. Report both, and label them.
Benchmark: Partial. Agency-side, the closest public data is placement speed: in Bullhorn GRID 2026 (n≈2,300, fielded Nov–Dec 2025), 56% of top performers place in under 10 days and 22% in three days or less.
3. Time to first submission
First candidate submitted − job order received (hours or days)
The agency speed metric every client feels and nobody publishes. It's the first proof you're working their job, and unlike time to fill it's entirely in your control.
Benchmark — read the caveat: None published. Track it per desk and per client, and watch the trend.
Conversion
4. Submission-to-interview ratio
Interviews ÷ submissions
The single best read on whether a recruiter understands the brief. A falling ratio usually means volume is standing in for judgment — or the client changed the job and nobody updated the spec.
Benchmark — read the caveat: No public benchmark exists. Benchmark against your own trailing twelve months, by desk and by client.
5. Interview-to-placement ratio
Placements ÷ client interviews
Where submission-to-interview tests your read of the brief, this tests your read of the candidate — and the client's hiring process. One client dragging this down is a client conversation, not a recruiter problem.
Benchmark — read the caveat: No public benchmark exists. Corporate funnel ratios measure application-to-interview in in-house hiring, which is a different thing.
6. Fill rate
Job orders filled ÷ job orders taken
The honest test of what you take on. A low fill rate is often a business development problem wearing a delivery costume: the desk is accepting job orders it was never going to fill.
Benchmark — read the caveat: No industry average is published. Bullhorn GRID 2026 found firms using AI throughout the workflow were 2x as likely to report fill rates above 75% — a survey association, not proof AI caused it. It implies 75% is a meaningful threshold. It isn't an average.
7. Offer acceptance rate
Offers accepted ÷ offers extended
Low acceptance usually traces back to something you could have found out earlier: a counteroffer, a salary gap, a second process.
Benchmark: No agency benchmark. Market context: Gartner reported offer acceptance fell to 51% in Q2 2025, from 74% in Q2 2023.
Money
8. Cost per hire (cost per placement)
(Total external costs + total internal costs) ÷ total hires in period
That's the commonly used form. For an agency it becomes cost per placement: job board spend, tooling and other direct costs are the external side, recruiter cost is the internal side. The calculator below runs exactly this formula on your inputs.
Benchmark: $5,475 nonexecutive, $35,879 executive (SHRM 2025 Benchmarking Reports) — employer costs, not agency costs. Separately, Appcast puts recruitment advertising cost per hire at about $851 (2024 data). That's ad spend only, a different number.
9. Gross margin / spread
Spread = bill rate − loaded pay rate · Margin = spread ÷ bill rate
The temp desk's real revenue line. Two desks with the same hours billed can be very different businesses once you look at spread.
Benchmark — read the caveat: The only public figure: 25.6% median gross margin for US IT temporary staffing in 2021 (Staffing Industry Analysts). One sector, five years old. Don't use it as your target.
10. Gross profit per recruiter
Gross profit in period ÷ FTE recruiters
The number that actually runs the business. Everything above it is a lever; this is the result. Your calculator output is your number only.
Benchmark — read the caveat: No public benchmark exists. The data sits inside ASA's members-only operations survey.
Retention and reuse
11. Redeployment rate
Contractors redeployed ÷ contracts ended
The cheapest placement you'll ever make is the contractor you already vetted, already paid to source, and who already trusts you. Every contract that ends without a next assignment is margin walking out the door.
Benchmark — read the caveat: No average published. 48% of agencies don't track it (StaffingHub 2026 State of Staffing, n=231). In Bullhorn GRID 2026, 85% of firms with redeployment plans place in under 20 days.
12. Client NPS / repeat business
% promoters − % detractors · repeat = orders from returning clients ÷ all orders
The lagging indicator that tells you whether the other eleven are working. Clients don't score you on ratios. They score you on speed and whether you understood what they needed.
Benchmark: Staffing client NPS 45 in 2024, up from 36 (2023) and 28 (2020) — the highest since ClearlyRated began benchmarking in 2010. Placed-talent NPS 30 (ClearlyRated).
A note on the time-to-fill figures you'll see quoted
SHRM's 39 days is a median across 4,600+ organisations. Greenhouse's Hire Standard reports 59.67 days for 2025 — a mean across 6,000+ North American companies on its platform. Employ's 2024 Recruiter Nation Report put the average at 41 days, from a survey of 1,200+ TA decision makers.
Three different measures, three different populations, all employers rather than agencies. A mean gets dragged up by a few roles that take forever; a median doesn't. None is wrong, and none is a target for your desk.
The 12 metrics on one page
Print it, pin it, argue about it on Monday.
| Metric | Formula | Who it's for | Benchmark | Source | Cadence |
|---|---|---|---|---|---|
| Time to fill | Offer accepted − job order opened (calendar days) | Client, owner | 39 days median, nonexec, 4,600+ orgs | SHRM 2026 | Monthly |
| Time to hire | Offer accepted − candidate entered process | Client, delivery manager | Partial: 56% of top performers place in under 10 days | Bullhorn GRID 2026 | Monthly |
| Time to first submission | First submission − job order received | Delivery manager, client | None published | — | Weekly |
| Submission-to-interview | Interviews ÷ submissions | Delivery manager, recruiter | None published | — | Weekly |
| Interview-to-placement | Placements ÷ client interviews | Delivery manager | None published | — | Monthly |
| Fill rate | Job orders filled ÷ job orders taken | Owner, delivery manager | No average; above 75% is a meaningful threshold | Bullhorn GRID 2026 | Monthly |
| Offer acceptance rate | Offers accepted ÷ offers extended | Delivery manager | No agency benchmark; market fell to 51% (Q2 2025) | Gartner 2025 | Monthly |
| Cost per placement | (External + internal costs) ÷ placements | Owner | Employer cost per hire $5,475 nonexec — not an agency figure | SHRM 2025 | Monthly |
| Gross margin / spread | (Bill − loaded pay) ÷ bill | Owner, temp desk lead | 25.6% median, US IT temp only, 2021 | SIA | Monthly |
| Gross profit per recruiter | Gross profit ÷ FTE recruiters | Owner | None public (ASA members only) | — | Monthly |
| Redeployment rate | Contractors redeployed ÷ contracts ended | Owner, temp desk lead | No average; 48% of agencies don't track it | StaffingHub 2026 | Monthly |
| Client NPS | % promoters − % detractors | Owner, account manager | 45 staffing industry (2024) | ClearlyRated | Quarterly |
SHRM and Gartner figures describe employers and candidates, not agencies. "None published" means we found no citable public benchmark from any association, analyst or vendor.
The Agency Benchmarks That Don't Exist
You've probably seen an "industry average" submission-to-interview ratio quoted with total confidence. Here's the uncomfortable thing: we went looking for the source behind the agency numbers owners most want, and for these there isn't a public one from any association, analyst or vendor:
- Submission-to-interview and submission-to-placement ratios
- Placements per recruiter
- Gross profit per recruiter
- An industry-average fill rate
- Offer acceptance rate for agency placements
- Permanent placement fee percentages
- Temp gross margin by sector, outside a single 2021 IT-sector figure
Two paywalls stand between agencies and their own benchmarks
The two datasets that would answer most agency benchmarking questions — ASA's Staffing Operations Benchmarking Survey (120+ benchmarks including gross margin, placement fees and recruiter productivity) and SIA's Gross Margin and Bill Rate Trends — are both members-only. That's why almost every "industry average" you'll see quoted in agency blogs has no source attached to it.
How to benchmark without industry data
Honestly? Your own history beats an industry average. An average blends healthcare temp with executive search with light industrial. It was never going to describe you. Instead:
- Trailing twelve months. Compare this month's ratios to your own rolling year. It irons out seasonality and tells you whether you're getting better.
- Cohort by desk. Your best desk is the benchmark for the others — same market, same tools, same brand.
- Cohort by client. If one client's interview-to-placement ratio is half everyone else's, the problem probably isn't your recruiters.
- Cohort by role type. Perm and temp, volume and specialist, never in the same average.
If you see a blog quoting an "industry average submission-to-interview ratio," ask for the source. There isn't one.
Cost Per Hire Calculator for Agencies (Cost Per Placement)
The common cost-per-hire formula is (external costs + internal costs) ÷ hires. This calculator is the same formula with agency inputs: job board spend, tooling and other direct costs are the external side; recruiter cost is the internal side; placements replace hires. Then it sets that cost against the gross profit each placement earns.
Two rules. Every output is your number, not a benchmark. And there's no published agency average for any of them to compare against — that's the point of the section above.
Cost-per-placement calculatorEnter one month of your own figures. Nothing is stored or sent anywhere.
Your numbers — not benchmarks
- Your total cost
- —
- Everything it cost to run the desk this month.
- Your cost per placement
- —
- Total cost spread across the placements you made.
- Your gross profit per placement, after cost
- —
- What each placement leaves once its share of cost is paid.
- Your net contribution
- —
- Gross profit earned minus total cost. Negative means the desk lost money this month.
- Your break-even placements per recruiter per month
- —
- How many placements each recruiter needs just to cover cost.
- Your gross profit per recruiter
- —
- Gross profit earned this month divided by FTE recruiters.
These are calculated only from what you entered. There is no published agency average for cost per placement or gross profit per recruiter to compare them against — the useful comparison is your own trailing twelve months.
The formulas, in full
total_cost = job_board_spend + tooling_cost + recruiter_cost + other_direct_costs cost_per_placement = total_cost / placements gross_profit_per_placement = avg_gross_profit - cost_per_placement net_contribution = (avg_gross_profit × placements) - total_cost break_even_placements_per_recruiter_per_month = (total_cost / recruiters_fte) / avg_gross_profit gross_profit_per_recruiter = (avg_gross_profit × placements) / recruiters_fte
Where a denominator is zero (no placements, no recruiters, or zero average gross profit), the output shows "—" instead of a number.
Cite this calculator: HRmony.ai, "Cost-per-placement calculator," in Martin Gordon, Recruitment Metrics for Staffing Agencies: The 12 That Actually Matter, October 14, 2026, https://hrmony.ai/blog/recruitment-metrics-staffing-agencies#cost-per-placement-calculator.
Retire the $4,700 figure
The widely quoted "$4,700 average cost per hire" comes from a 2022 SHRM news article that names no underlying report or data year. Older still is "$4,100 and 42 days," which is from 2016 and still circulating. SHRM's current published figures are $5,475 for nonexecutive roles and $35,879 for executive roles (2025 Benchmarking Reports). And remember: both are employer costs, not agency costs.
Metric Theatre: The Numbers That Look Like Rigour
Calls made. CVs sent. Candidates added. LinkedIn messages. Job orders taken. They fill a leaderboard beautifully. They're also the easiest numbers in the building to inflate.
Reward CVs sent and you get more CVs, not better ones — and a submission-to-interview ratio that quietly falls apart. Reward job orders taken and you get roles nobody was going to fill. The activity was real. It just wasn't the point.
The one-line test for any metric on your dashboard: does this number change what we do next week? If not, it's theatre.
What to Put on a Weekly Recruitment Metrics Dashboard
Six numbers per role. Not twelve, and certainly not 23. Each role sees the numbers it can actually move.
| Role | Six numbers | Why | Where it comes from |
|---|---|---|---|
| Agency owner |
| These six say whether the business is making money and whether clients will keep sending orders. | Finance for cost and GP; ATS/CRM for placements, job orders and contract end dates |
| Delivery manager |
| These show where the funnel is leaking this week, and whether the leak is a recruiter or a client. | ATS pipeline stages and job order records |
| Individual recruiter |
| Every one of these changes what the recruiter does on Monday morning. | ATS pipeline, contract end dates, and the break-even output from the calculator |
Notice the last column. Most of these numbers live in a CRM (job orders, client history) or an ATS (pipeline, contract dates). If those are separate systems, every Monday starts with someone reconciling exports — the practical case for running ATS and CRM as one system. Picking from the analytics and reporting category matters less than where the data lives.
Running a volume desk? See the high-volume hiring playbook. For which of these numbers automation actually moves, read the guide to recruitment automation for staffing agencies covers the ROI side.
Your redeployment rate is hiding in your own database
HRmony's Pulse ranks every candidate already in your ATS against any open job order in under two minutes — so the contractor finishing on Friday shows up on Monday's shortlist, not in a spreadsheet nobody opens. And because ATS and CRM are one native base, every number on the dashboard above comes from one place. See what's included on the pricing page.
See How It Works →FAQ: Recruitment Metrics for Staffing Agencies
What are the most important recruitment metrics for a staffing agency?+
What's the difference between time to fill and time to hire?+
How do you calculate cost per hire?+
What is a good fill rate for a staffing agency?+
What is redeployment rate and how do you calculate it?+
What's a good time to fill in 2026?+
What is quality of hire and how do you measure it?+
Where can I find staffing industry benchmarks?+
Further Reading
On HRmony
Martin Gordon
Co-founder & CEO, HRmony.ai
Martin has 15+ years in recruiting technology and talent operations. He co-founded HRmony.ai so agency owners could see every number that matters — candidates, clients and placements — in one system instead of three exports. About HRmony
One platform. ATS + CRM + AI — built for agencies.
Put your job orders, candidates and placements in one place and the metrics stop being a Monday-morning project.
Try HRmony →