The 2026 Staffing Industry Loyalty & Referral Benchmark
Across three published verticals, referred candidates stayed 50% to 82% longer than candidates from job boards. They also redeployed more often and started faster in light industrial. Based on data from 882,004 placements across two years.
About this benchmark. Data is drawn from placed candidates across staffing agencies running referral programs on the Staffing Referrals platform. Findings reflect patterns measured across this network. An external anchor, Harvard Business Review (Gautier & Munasinghe, 2020), finds referred candidates stay longer and perform better. This benchmark quantifies how much longer, specifically for the staffing industry. See the data note at the end for sample size and methodology.
In light industrial, referred candidates work nearly twice as long as job-board candidates.
A typical referred light industrial worker stays 64 days. A typical job-board worker stays 36 days. At the high end, the difference is larger still. The top quarter of referred workers cross 166 days, against 109 for the top quarter of job-board workers. This pattern repeats across every vertical with sufficient data to publish.
"Referred candidates are of higher quality than applicants from the general public and are more likely both to receive and accept an offer, stay at the job longer, and perform better."
Harvard Business Review · Gautier & Munasinghe · "Build a Stronger Employee Referral Program" · 2020
Prior research finds this pattern in general employment. This benchmark report tests whether it holds in staffing. Referred candidates worked 50-82% more lifetime days across the cohorts analyzed.
Referred candidates stay longer across all three published verticals.
Referred candidates work more days than candidates from any other source, and it isn't a small difference. In light industrial the gap vs. major job boards is 78%. In healthcare it's 82%. In travel nursing it's 50%. A third source, candidates applying through agency career websites, performs nearly identically to major job boards. Posting on your own site produces the same tenure outcome as a major job board.
Agency career sites and major job boards produce the same tenure outcome. Referral programs do not.
Bars scaled within each vertical. Not cross-vertical.
Why the healthcare aggregate (180d) sits above the travel nursing sub-cut (174d): the per-diem sub-cohort is asymmetric across sources. Referred per-diem candidates accumulate many shifts over a long relationship; open-marketplace per-diem candidates churn early. Travel nursing (+50%) is the cleaner single-business-model headline.
Median lifetime days worked per candidate, 2024-2025. Agency website classification covers source strings matching website, career site, company website, and direct-apply variants across contributing ATS systems. Anonymity floor cleared for all cells: ≥22 agencies per bucket. Source: dataset, tenure analysis, 2024-2025.
Referred candidates come back for second and third assignments.
Because referred candidates keep coming back, they work more days than job board candidates. Referred travel nurses average 2.4 assignments; job board nurses average 1.4. However, each individual assignment is similar in length (referred contracts are actually a touch shorter on average). Repeat business builds the longer career. Single contracts don't get longer. It's the same in light industrial with a smaller repeat effect, plus a notable per-assignment lift on top.
Referred TN candidates average 2.40 assignments (approaching two 13-week contracts); open-marketplace 1.45 (~one and a third). Per-assignment duration is comparable (174÷2.40 ≈ 73 days vs. 116÷1.45 ≈ 80 days; referred ~9% shorter). The lifetime tenure lift is a redeployment story. Contract length stays roughly the same.
LI shows both effects: +15% redeployment lift and +55% per-assignment duration (64÷1.84 ≈ 35 days vs. 36÷1.60 ≈ 22 days). The two combine to the +78% lifetime tenure lift in Section 01.
Mean assignments per acquired candidate, 2024-2025. Per-assignment durations are derived (median lifetime days ÷ mean assignments per candidate); they approximate but don't exactly equal the per-row median, because lifetime days are measured as a median while assignments per candidate are measured as a mean. Healthcare aggregate is excluded from this chart because per-diem cadence (open-marketplace per-diem averages 10+ single-shift placements) makes the count incomparable. Source: dataset, tenure analysis. Active assignments included (capped at 730 days).
In light industrial, referrals get to work three days faster.
Referred light industrial candidates reach their first paid day in 14 days. Job-board candidates take 17. In a cohort where most placements close within 30 days, three days is real money. It's roughly 18% of the time-to-revenue window. Healthcare moves in the opposite direction. Referred clinical candidates take longer to place.
In healthcare, referred candidates take longer to reach their first paid day than open-marketplace candidates. The likely driver is a selection effect. Referred clinical candidates hold out for higher-paying contracts. Operators should treat this as a real bench time line item. It deserves more than a footnote.
Light industrial cohort only (cleanest speed comparison). Median days from candidate creation to first placement, 2024-2025. Source: Q19 Section A. See Methodology "Anticipated questions" for healthcare speed detail.
Where the lift comes from: who refers, how often, and how the network concentrates.
A small group of ambassadors does most of the work.
Across the dataset, 83,985 ambassadors (referrers) sent in at least one referral. Together they produced 174,900 referrals and roughly an 11% network conversion rate. The work isn't evenly distributed though. The top 10% of ambassadors generate 53% of all referrals. That's a heavy concentration, and the obvious instinct is to chase the heavy producers harder. The next section shows why that instinct is partly wrong.
83,985 ambassadors submitted at least one referral in 2024-2025. Top 10% = top 8,399 ranked by referrals submitted. Source: dataset, ambassador economics.
High-volume referrers aren't always the best-converting referrers.
A third of all ambassadors send in more than one referral, which is a good sign of a healthy program. Conversion declines as referral volume rises: 14% for one-time referrers vs. 6% for ambassadors with 25+ referrals. That trend may reflect vertical mix (for example, clinical roles convert at 7% regardless of referral volume). The better first test is asking one-time referrers for a second referral.
Distribution of ambassadors by total referrals submitted. Multi-referrer rate (2+ refs) = 32%. Source: dataset, ambassador economics, 2024-2025.
How often a referral becomes a placement depends almost entirely on the vertical.
In light industrial, 14% of submitted referrals end up working. In tech / IT, the number is 2.7%. The 5x spread isn't only about referral quality. It reflects how each vertical actually hires: credentialing in healthcare, long evaluation cycles in technical roles, and fast cycles in industrial. A program built around tech / IT placement rates will look like it's failing when it's just running on a slower clock.
Placement rate = placed referrals / total referrals submitted, 2024-2025 data window. Source: dataset, ambassador economics. Differences across verticals reflect the friction of each vertical's evaluation cycle. Source quality isn't the driver.
Three things this data made us reconsider.
These aren't prescriptions. They're observations that surprised us, and they may not match how referral programs get evaluated today.
A referral isn't a hire. It's a candidate relationship.
Most referral programs are evaluated on a single-placement basis. Did this referral convert, and what did the bonus cost? A better way is to measure the full candidate relationship. Referred candidates in travel nursing average 2.40 placements vs. 1.45 for job-board hires. In light industrial, 1.84 vs. 1.60. Each of those placements carries its own gross profit. The economics of a referral only make sense when you run them over the full candidate lifetime, calculating the conversion rate x gross profit per placement x average placements per candidate. That number is materially different from what a single-hire calculation produces, and it changes the true value of a referral bonus.
The person most likely to refer again already referred once.
Conversion holds near 13% for ambassadors with two to four referrals, almost identical to the first referral at 14%. The ambassador who already referred once is the most likely source of the next placement. Most programs spend their activation budget finding new ambassadors. The data suggests the better first question is whether every one-time referrer has been asked for a second name.
Conversion rate is the wrong number to compare verticals.
Light industrial converts at 14%. Tech / IT at 2.7%. That looks like a five-to-one advantage for industrial. However, gross profit per placement varies by several multiples, and referred candidates redeploy at different rates by vertical. The right comparison is lifetime referral value. That calculation will look different at every agency, and the vertical ranking it produces likely isn't the same as the conversion-rate ranking.
Published by StaffingHub · staffinghub.com · Questions or custom cuts: [email protected]
About the data.
All findings come from placed candidates across contributing staffing agencies, covering the 2024-2025 period. The headline metric is median lifetime days worked per candidate: the total days across all placements a candidate completed with the same agency. Three sources are compared: referred candidates (sourced through an agency's referral program), candidates who applied via the agency's own career website, and candidates placed via major job boards. Dataset is limited to agencies operating an active referral program; reported lift may reflect both referral quality and program maturity.