The Police Recruiting Funnel: What Each Stage Should Convert At

The useful part of that number is not 1.4%. It is that almost no agency in the country can produce the same figure for itself.

Ask a chief how many officers they are down and the answer comes back immediately. Ask how many applications it takes to produce one sworn officer, and where in the process the other applicants went, and the room goes quiet. The data exists somewhere — in an applicant tracking system, a background unit's spreadsheet, a recruiting sergeant's notebook — but nobody has ever put it end to end.

This page assembles what has actually been published, stage by stage, with a citation on every number. It is the reference we wanted and could not find.

Why nobody can tell you what normal looks like

Start with the uncomfortable part: the stage-by-stage benchmark tables circulating in this industry are not sourced. They typically credit BJS, IACP or PERF. We checked all three.

• The Bureau of Justice Statistics' most recent hiring and retention tables report 64,200 sworn vacancies, 55,000 hires and 57,400 separations. They contain no applications-received count, no stage pass rates and no conversion rates of any kind.¹
• PERF's annual staffing survey collects four things: sworn headcount, hires, resignations, retirements. Not applications.²
• IACP's 2024 survey asked 1,158 agencies what they perceive about recruiting difficulty. It is a survey of opinion, and a good one, but it does not measure a pipeline.³
• GAO's February 2026 review of law enforcement recruitment and retention covers resignations, retirements and retention strategies. No pipeline metrics.⁴

No federal agency, professional association or academic body publishes police hiring conversion data as a national series. A national applicants-per-hire figure does not exist. Any table that presents one as a settled fact invented it.

What does exist is a handful of city audits and oversight reports where somebody was forced to count. Those are worth more than a synthesized average, because they are real agencies with real numbers attached to a public document.

The four funnels that have been published

Atlanta and Nashville figures come from municipal audit and oversight reports; San Francisco from news reporting on SFPD data rather than an audit; Los Angeles from RAND's organizational assessment. Nashville and Los Angeles published stage detail but no end-to-end conversion figure. Note: RAND states LAPD's 2024 application count as both 13,407 and 13,047 in different passages; we use 13,047 here because it is the figure RAND pairs with its testing count.

Two agencies two thousand miles apart, on different systems, land in the same neighbourhood: roughly one to one and a half percent, application to sworn. Read the endpoints carefully, because they differ — Atlanta's figure counts officers hired, San Francisco's counts academy graduates. Applying Atlanta's own 79% graduation rate to its hires puts it nearer 1.1% measured at graduation. Treat one in seventy to one in ninety as the plausible band, not a point estimate. If your agency's number is dramatically better, check the denominator — you are probably counting qualified applicants rather than applications.

Stage by stage

Application started to application complete

Atlanta: of 12,129 applications entered, 35% were incomplete or did not meet minimum requirements. 7,864 complete applications, or 64.8%.

A low number here is rarely a candidate-quality problem. It is a form problem, a minimum-qualifications-clarity problem, or an abandonment problem — someone started, hit a question they could not answer, and nobody ever called them.

Complete application to appearing for testing

This is where the published data gets brutal.

Nashville invited 9,617 applicants to civil service testing. 2,717 showed up. 28%. Seven of every ten invited candidates never appeared for the first step.

Los Angeles, in calendar year 2024: 7,566 multiple-choice tests completed against 13,047 applications received. 58%.

Nashville's oversight board is explicit that these candidates did not show up — nobody was disqualified and nobody failed anything. They applied, waited, and stopped. Los Angeles does not break its gap down, and a calendar-year ratio of tests to applications is not a true cohort measurement, so read the LAPD figure as an indicator rather than a conversion rate.

Testing pass rate

Nashville: 2,277 of 2,717 test-takers passed. 84%.

Read that against the line above it. Of everyone invited to test, 72% never came — and of those who did, five in six passed. Testing is not where agencies lose people. Showing up is. Agencies that respond to a staffing shortfall by re-examining their written exam are usually working on the wrong stage.

Background investigation

The spread here is enormous, and the spread is the finding.

Nashville's own two figures — 79% in a single year, 56% pooled across five — differ by 23 points inside one agency. That alone is why a single "healthy background pass rate" does not exist, and why the Atlanta row above is labelled for the wider span it actually covers. What matters is whether your rate is stable, and whether your disqualification reasons cluster. If one reason accounts for a large share of your background failures, you do not have a background problem. You have a sourcing problem that background is paying for.

Background is also the stage where time does the damage. Chicago's Office of Inspector General found 2,400 applicants from the 2016–2018 entrance exams still sitting at background investigation as of May 2021, and 4,020 applications from those pools that never advanced past the stage at all. Nearly 60% of applications that terminated at background had not received a formal rejection notice as of spring 2021 — three to six years after the candidate applied.⁹

Background investigations are labour-intensive and genuinely constrained. That is a real operational limit, not a failure of will. But a candidate who sits in an unmanaged queue for a year is lost whether or not they would have passed.

Conditional offer to academy seat

No agency in the published set separates this stage. It is the one candidates most often exit for a competing offer, and almost nobody measures it.

Academy

This is the only stage with solid national data. In 2022, 747 state and local academies trained 60,214 recruits and 51,144 completed — 84.9%. BJS reports the non-completions as more than 14% of all recruits: 8% involuntarily, 5% by voluntary withdrawal, about 1% other or unknown.¹⁰

The national average conceals a 14-point spread by academy model:

Agency-level figures sit in the same band: Atlanta averaged 79% graduation across its 2017–2020 classes; RAND put LAPD academy attrition at 14.75% for 2022–2024.

The 8%/5% involuntary-versus-voluntary split is the number to pull for your own classes, because the two have opposite fixes. Involuntary attrition points back at selection. Voluntary attrition points at the gap between what recruiting promised and what the recruit found.

Field training

There is no published national field training completion rate. The quantitative work that exists is single-agency, the dedicated national study of field training programs is qualitative, and agency hiring audits stop at academy graduation. If your agency tracks field training completion, you hold data almost nobody else has.

The pattern: most of your losses happen before anyone screens anybody

Lay the stages side by side and the shape is consistent across every published funnel.

Atlanta's single largest loss was not background and not the academy. Of 7,864 candidates with complete applications, 91.5% never finished the waiver packet — the personal-history paperwork that opens a background investigation. 7,197 people. Separately, the auditor found that 92% of applicants who began the packet failed to finish it.

Nashville's largest loss was 72% of invited candidates not appearing for a test.

Los Angeles lost 42% of applicants before the multiple-choice exam.

This changes the diagnosis. An agency that believes it has a standards problem, or a candidate-quality problem, or an applicant-volume problem, most likely has an abandonment problem — and abandonment is a follow-up and cycle-time issue, not a hiring-criteria issue. It is also the cheapest thing in the entire funnel to fix, because those candidates already chose you once.

It fits what agencies report about their own process. IACP's 2024 respondents said only a small number of candidates withdraw formally, and that once formal testing begins, eliminations cluster at background investigation and the in-person interview. Both can be true. Candidates rarely withdraw on the record. They go quiet, and the file ages out.

Volume came back. Conversion did not.

If your recruiting strategy was designed in 2021, it is now aimed at a problem that has partly resolved itself.

PERF's May 2026 survey of 197 agencies representing more than 138,000 sworn officers found total staffing up 1.4% in 2025, back to where it stood in 2022. Responding agencies hired 17.6% more officers in 2025 than in 2024, and 40% more than in 2021. Resignations were 15% below their 2021 level.²

Application volume followed the same curve. LAPD went from 8,408 applications in 2022 to 13,047 in 2024, a 55.2% increase.⁸

And yet: in IACP's summer 2024 survey, responding agencies were staffed at approximately 91% of authorized strength³ — a figure GAO later cited⁴ — and as of March 2025 LAPD's average time from application to hire was 349 days.⁸

Run your own numbers

You do not need a platform to do this once. You need somebody to sit down for a day with whatever system holds your applicants and count.

1. Pick a closed cohort. Every application received in a single quarter, at least 12–18 months ago, so the cohort has had time to finish. Not a snapshot of who is in the pipeline now — a snapshot flatters you, because your failures have already left it.
2. Count the survivors at each stage. Complete application, appeared for testing, passed testing, entered background, cleared background, received conditional offer, started academy, graduated, completed field training, still employed today.
3. Divide each stage by the one before it. That is your stage conversion. Divide the final number by the original applications for your end-to-end rate.
4. Find your largest single drop. One stage will be materially worse than the rest. That is where your next dollar goes.
5. Split each loss into disqualified, withdrew and went dark. This is the step almost every agency skips, and it is the one that tells you what to fix. Disqualified means your standards worked. Went dark means your process lost someone you had already won.
6. Record days in stage. Median, not average — a handful of ancient files will drag an average past usefulness.

Do it once and you have a baseline. The difficulty is doing it every month, for every recruiter, for every lead source, without a person rebuilding the spreadsheet each time.

What a shortfall at each stage tells you

One caution on all of it. Every stage rate above is a ratio, and a ratio moves when either number moves. An agency that tightens its top of funnel will watch its background clear rate rise and its end-to-end conversion improve while producing the same number of officers. Read conversion alongside absolute hires, always.