Back to blog
Hiring Benchmarks13 min read

Which Hiring Sources Produce the Longest Tenure? Retention by Source of Hire in 2026

K
Klearskill TeamNovember 18, 2026

Referral hires stay with their employers for a median of 38 months, compared with 22 months for non-referral hires, according to EmployeeReferrals.com data published by iCIMS covering 91,000 employees. That 70% gap is the clearest signal in retention by source of hire data, but it is not the whole story. This guide sets out which hiring sources produce the longest tenure, how to measure it in your own business, and the benchmarks worth holding yourself to.

Quick Answer

Retention by source of hire measures how long employees stay based on the channel that brought them in. Internal moves and employee referrals consistently produce the longest tenure: referral hires stay around 70% longer than non-referral hires, and companies that regularly hire internally keep employees 41% longer. Job boards and agencies vary far more by screening quality.

What Is Retention by Source of Hire?

Retention by source of hire is the percentage of employees still employed after a set period, usually 90 days, 12 months or 24 months, grouped by the channel through which they were recruited. It answers a simple but powerful question: which of our hiring channels brings in people who stay?

The metric has three components. The first is the source, meaning the channel credited with the hire, such as an employee referral, an internal move, a job board, the careers site, a recruitment agency, direct sourcing on LinkedIn, or a graduate programme. The second is the cohort, meaning everyone hired from that source within a defined window, typically a quarter or a year. The third is the retention window, the point at which you check whether each person is still employed.

The inputs are your applicant tracking system's source field and your HRIS's employment records. The output is a retention rate per source, often alongside median tenure, so you can compare channels on equal terms.

It is important to be clear about what this metric is not. It is not the same as source of hire volume, which simply counts how many hires each channel produced. A job board can deliver 40% of your hires and still be your weakest channel if those hires leave within a year. It is also not quality of hire, although the two are closely related; quality of hire usually blends retention with performance ratings and hiring manager satisfaction. Finally, it is not overall employee turnover. Overall turnover tells you how many people leave; retention by source of hire tells you where the leavers came from, which is far more actionable for a recruitment team.

Why Retention by Source of Hire Matters

Early attrition is expensive, and it is more common than most businesses admit. The CIPD's 2024 recruitment and retention research with Omni RMS, based on a survey of over 1,000 HR professionals, found that 41% of respondents said new recruits always, mostly or sometimes quit within the first 12 weeks. A further 27% said selected candidates sometimes failed to turn up on their first day at all.

Each of those departures carries a real cost. Gallup estimates that replacing an employee costs between one-half and two times their annual salary, and puts the annual cost of voluntary turnover to US businesses at around $1 trillion. Gallup also found that 52% of voluntary leavers said their manager or organisation could have done something to prevent them leaving, which suggests a large share of early attrition is avoidable.

Source data shows where to intervene. If referral hires and internal moves stay longest, then investing in a referral programme and internal mobility is a retention strategy, not only a recruitment one. LinkedIn's Global Talent Trends research found that employees stay 41% longer at companies that regularly hire from within, and the same research cited a Gartner estimate that turnover driven by a lack of future career opportunities costs average-sized companies $49 million per year.

Without source-level data, recruitment budgets tend to follow volume. With it, they follow value.

How Retention by Source of Hire Works

Measuring retention by source of hire is a mechanical process once the data is clean. It works in five steps.

Step one: capture the source at application. Every candidate record needs a single, standardised source value. Use a fixed list rather than free text, and decide in advance how to credit candidates who touch several channels, for example first touch or the channel that produced the application.

Step two: carry the source through to the employee record. When a candidate becomes an employee, the source field must move from the applicant tracking system into the HRIS. This is where most organisations lose the data, because the two systems are not connected or the field is not mapped.

Step three: define cohorts. Group hires by source and by start date, typically by quarter. Mixing a January cohort with a December cohort distorts results because the later group has had less time to leave.

Step four: check status at fixed windows. For each cohort, record how many people are still employed at 90 days, 12 months and 24 months. Separate voluntary from involuntary leavers if you can; a channel producing many dismissals tells a different story from one producing many resignations.

Step five: compare and act. Rank sources by retention rate and median tenure, then look at cost per hire alongside. A cheap channel with poor retention can be more expensive in practice than a costly channel whose hires stay for years.

The underlying reason the differences exist is information quality. Referrers and internal managers know both the role and the person, so candidates arrive with realistic expectations. Channels with less pre-screening rely far more on your own assessment to filter out poor fits.

How to Measure Retention by Source of Hire

The core formula is straightforward:

Retention rate by source = (hires from source still employed at the end of the window ÷ total hires from source who started in the cohort) × 100

For example, if you hired 40 people through job boards in the first quarter of 2025 and 26 were still employed 12 months later, your 12-month job board retention rate is 65%. If 18 of 20 referral hires from the same quarter remained, your referral retention rate is 90%.

Pair the rate with median tenure for leavers and stayers combined, because averages are easily distorted by a handful of very long-serving employees. The iCIMS analysis noted above uses medians for exactly this reason.

As a working target for a 12-month window, best-in-class teams retain 90% or more of referral and internal hires after 12 months, and keep the gap between their strongest and weakest external channel under 15 percentage points. Average teams sit around 75% to 80% retention at 12 months across all sources, with a gap of 20 points or more between referrals and their weakest channel. Treat these as internal targets rather than universal laws: sector, role level and labour market conditions all move the numbers.

Report the metric quarterly, but make decisions annually. Small cohorts produce noisy results, and a single leaver can swing a 10-person cohort by 10 points.

Common Retention by Source of Hire Mistakes

Letting candidates choose their own source

Self-reported "how did you hear about us" fields are unreliable, because candidates often name the last place they saw the job rather than the channel that prompted them to apply. Fix this by recording source automatically through tracked links and referral codes, and use self-reporting only as a fallback.

Lumping all job boards together

A general job board and a specialist niche board can produce very different retention, yet many businesses group them as one source. Track each board separately for at least a year, then merge only those whose results are genuinely similar.

Measuring too early

A 90-day snapshot captures early regret but misses people who leave later in their first or second year. Track at 90 days, 12 months and 24 months, and avoid drawing conclusions about a channel until at least one full 12-month cohort has matured.

Ignoring screening quality

Poor retention from high-volume channels often reflects rushed screening rather than a bad channel. When a recruiter reviews 300 CVs in an afternoon, mismatched expectations slip through. Apply the same structured criteria to every applicant, regardless of source, so the comparison reflects the channel rather than how tired the reviewer was.

Treating referrals as a guaranteed win

Referral programmes can also narrow diversity if employees mainly refer people like themselves. Monitor the demographic mix of referral hires and screen referred candidates against the same criteria as everyone else, so the retention benefit does not come at the expense of fairness.

Retention by Source of Hire Benchmarks

These benchmark statements summarise what the strongest available data shows about which hiring sources produce the longest tenure.

  • Employee referral hires stay a median of 38 months, compared with 22 months for non-referral hires, a 70% difference, based on EmployeeReferrals.com data from over 50 companies and 91,000 employees reported by iCIMS.
  • In high-turnover companies, referral hires stay 122% longer than non-referral hires; in low-turnover companies, they stay 51% longer, according to the same dataset.
  • Employees stay 41% longer at companies that regularly hire from within, according to LinkedIn's Global Talent Trends research.
  • Employees who submit referrals stay around 20% longer than standard hires, so a referral programme improves retention among existing staff as well as new hires.
  • Referrals accounted for more than 30% of all hires and 45% of internal hires in SilkRoad's Sources of Hire analysis of 329,000 hires, as reported by SHRM.
  • 41% of HR professionals surveyed by the CIPD report that new recruits at least sometimes quit within 12 weeks which makes the 90-day window the first retention checkpoint every business should track.

What Drives the Differences Between Sources

The tenure gap between sources is not magic. It comes down to three factors: expectation accuracy, cultural context and the depth of assessment before an offer.

Referral candidates usually hear an honest account of the job from someone who already does similar work. They know the hours, the manager's style and the frustrations before they apply. Internal candidates know even more. That reduces the "this is not what I signed up for" departures that dominate early attrition.

External channels such as job boards and agencies start with far less shared context. The candidate relies on a job advert, and the employer relies on a CV. McKinsey's research on the Great Attrition found that 35% of people who left their jobs listed uncaring leaders among their top three reasons, alongside inadequate pay, limited career advancement and a lack of meaningful work. Most of those factors could be surfaced earlier with better screening and more transparent job descriptions.

The practical lesson is that external channels can close much of the gap. Detailed, honest job adverts, structured screening against clear criteria and realistic interview conversations bring the expectation accuracy of a referral to an applicant who has never met your team. Skills development matters too: LinkedIn found that 73% of employees say they would stay longer at their company if there were more skill-building opportunities.

How to Improve Retention From Every Source

Start by fixing your data. If your ATS and HRIS do not share a consistent source field, no benchmark will be reliable. Spend the first month cleaning historic source data for at least the last 24 months of hires.

Next, double down on what works. Formalise your referral programme with clear rewards, fast feedback to referrers and the same assessment standard for every referred candidate. Publish internal vacancies before external ones, giving existing employees a genuine first look.

Then raise the floor on high-volume channels. The biggest retention gains usually come from improving screening on job boards and the careers site, because that is where most applications arrive and where rushed manual reviews do the most damage. Consistent, criteria-based screening of every CV, rather than a skim of the first 50, means the shortlist reflects the job's real requirements.

Agencies and graduate programmes deserve their own review. Agency hires are often placed quickly into hard-to-fill roles, so judge them against the difficulty of the vacancy rather than against referrals alone, and check whether your agency terms include a rebate period that reflects your real early attrition. Graduate and apprentice cohorts tend to have structured development built in, which supports retention, but they also face natural career moves after two or three years. Track them over a longer window, and compare them with each other year on year rather than with experienced hires.

Finally, close the loop with hiring managers. Share retention by source of hire results every quarter, and ask managers which hires surprised them, positively or negatively. Their observations often explain the numbers faster than any dashboard.

Frequently Asked Questions

What is retention by source of hire?

Retention by source of hire is the percentage of employees still employed after a defined period, grouped by the channel that recruited them, such as referrals, job boards or agencies. It shows which channels bring in people who stay, so recruitment budgets can follow long-term value rather than application volume alone.

Which source of hire has the best retention?

Internal moves and employee referrals consistently show the best retention. Referral hires stay a median of 38 months compared with 22 months for non-referral hires, according to EmployeeReferrals.com data reported by iCIMS. LinkedIn research shows companies that regularly hire internally keep employees 41% longer than those that do not.

How do you calculate retention by source of hire?

Divide the number of hires from a source who are still employed at the end of your chosen window by the total number of hires from that source in the cohort, then multiply by 100. Use fixed windows such as 90 days, 12 months and 24 months, and compare cohorts that started in the same period.

What is a good 12-month retention rate for new hires?

A sensible target is 90% or more for referral and internal hires after 12 months, with an overall rate across all sources of around 80% or higher. Results vary by sector, role level and location, so compare against your own historic data first and industry figures second.

Why do job board hires often leave sooner?

Job board candidates usually know less about the role and the team before applying, so mismatched expectations surface after they start. High application volumes also encourage rushed manual screening. Clear job adverts and consistent, criteria-based screening of every CV close much of the gap with referral hires.

How long should I track retention by source?

Track each cohort for at least 24 months, with checkpoints at 90 days, 12 months and 24 months. The 90-day point captures early regret, whilst the later checkpoints reveal whether a channel produces people who build lasting careers. Avoid judging a channel until a full 12-month cohort has matured.

Stop Screening CVs Manually in 2026

Klearskill screens every CV against your criteria with 97% accuracy and cuts screening time by 92%, so high-volume channels deliver hires who stay. Two plans, both unlimited on jobs and candidates: Starter at $10/month with your own AI key, or Pro at $50/month with AI included. Start with Klearskill free on 1 job and 25 CVs, no card needed.

RetentionSource of HireHiring BenchmarksRecruiting Metrics

Screen smarter, hire faster

Put these ideas into practice with AI-powered CV screening built for modern hiring teams.