Retention programmes are usually built around the people who have been there a while: the flight risks on the succession chart, the ones with a competing offer, the ones whose engagement scores dropped. It is a reasonable place to look and it is not where the risk is densest.
The finding
Rubenstein, Eberly, Lee and Mitchell's 2018 meta-analysis of the antecedents of voluntary employee turnover pooled 117 studies on the relationship between tenure and quitting. The correlation is −0.27.
Negative means: the longer someone has been there, the less likely they are to leave. And 0.27 is a substantial correlation by the standards of this literature, where the strongest individual predictors sit around −0.29 for organisational commitment, −0.23 for job security and −0.20 for engagement. Tenure is in that top group, and unlike the others it requires no survey to measure.
The uncomfortable arithmetic follows immediately. The costs of a hire, advertising, agency fees, interviewing time, onboarding, and the weeks before the person produces anything, are all paid at the start. The probability of losing them is also highest at the start. Every organisation is therefore paying its highest acquisition costs against its highest attrition risk, and most of them are looking somewhere else entirely.
What actually moves it
The best evidence on what changes a newcomer's trajectory is Bauer and colleagues' 2007 meta-analysis, which pooled 70 unique samples of newcomers and tested a model rather than a list.
The model is worth stating precisely, because the structure is the finding. Organisational socialisation tactics and the newcomer's own information seeking do not act directly on the outcomes. They act on adjustment, and adjustment has three components: role clarity, self-efficacy and social acceptance. Adjustment then mediates the effects on job satisfaction, organisational commitment, job performance, intentions to remain and turnover.
In plain terms, a new employee stays when they know what they are supposed to do, believe they can do it, and feel the people around them have accepted them. Everything an onboarding programme does reaches retention through one of those three or not at all.
A note on what this article claims. The Bauer paper is behind a paywall and only the abstract could be read directly, so this piece reports the model and its components and does not quote effect sizes for the individual paths. The structure is what the abstract establishes and the structure is the useful part.
Measured against what onboarding usually is
Take the three components and hold a typical first-week programme against them.
Role clarity. Rarely addressed directly. A job description written to attract applicants is not a description of what the person will actually do, and the gap between the two is discovered over weeks by inference. This is the cheapest of the three to fix and the least often attempted: a written account of what the role is accountable for, what it decides, and what "doing this well" looks like in ninety days.
Self-efficacy. Onboarding often works against it. A new starter is shown systems, policies and org charts at a volume nobody could absorb, which teaches them how much they do not know. What builds the belief that they can do the job is doing a small piece of it early and having it land.
Social acceptance. The one most damaged by hybrid work, and the one most often delegated to a buddy scheme with no structure behind it. A new person's read on whether they are accepted comes from whether their contributions get taken up, not from whether someone was assigned to have coffee with them.
The first-week compliance module, the equipment, the handbook: necessary, and none of them touches any of the three.
Where the manager sits in this
All three components are largely delivered by one person, and it is not anyone in HR.
Role clarity is a conversation with the line manager. Early wins are assigned by the line manager. Social acceptance is substantially a function of whether the manager visibly treats the newcomer's contributions as real. Onboarding is a managerial task with an administrative wrapper, and it is usually resourced the other way round.
That is worth holding next to what is happening to that population. Manager engagement fell five points in a single year, against one point for the workforce overall, and the meta-analytic evidence puts the performance consequences of losing a manager at roughly three times those of losing an average employee. An onboarding programme that depends on line managers having spare attention is depending on the scarcest thing in the building.
The same evidence says that development spend works hardest at exactly that level: programmes delivered to junior leaders produce significantly stronger behaviour change than those delivered further up.
What this changes in practice
Measure turnover by tenure band, not as one number. An organisation reporting 14 percent is reporting an average over a population whose risk varies by a factor that this literature puts at a correlation of −0.27. First-year attrition, second-year attrition and everything beyond are three different figures with three different causes.
Put the retention conversation before the six-month mark, not at the annual review. By the review the decision has usually been taken, and the exit interview will record a reason that was formed months earlier.
Judge onboarding against the three components. Not against completion rates or satisfaction scores, which measure whether the programme ran. Whether someone can say what they are accountable for, has already done something that worked, and has had a contribution taken up: those are answerable at thirty days and they are what the evidence says predicts whether the person is still there.
And do the costing properly while you are at it. A first-year departure is the most expensive kind, because the acquisition cost has been paid and none of it has been recovered, which is the argument set out in what losing an employee actually costs.
Common questions
When are employees most likely to leave? Early. A meta-analysis pooling 117 studies reports a correlation of −0.27 between tenure and voluntary turnover, meaning the probability of quitting falls as time with the employer rises.
How strong a predictor of turnover is tenure? Comparable to the strongest attitudinal predictors. In the same body of work, organisational commitment sits at −0.29, job security at −0.23 and engagement at −0.20. Tenure has the practical advantage of needing no survey to measure.
What makes onboarding work? A meta-analysis of 70 newcomer samples finds that socialisation tactics and information seeking affect outcomes through newcomer adjustment, which has three components: role clarity, self-efficacy and social acceptance. Everything else reaches retention through one of those or not at all.
Is a longer onboarding programme better? The evidence addresses content rather than duration. A short programme that establishes what the person is accountable for, gets them an early win and makes their contributions visible is doing the three things that matter; a long one that does none of them is not.
Who is responsible for onboarding? On this evidence, the line manager. Role clarity, early wins and social acceptance are all delivered by the person the newcomer reports to, whatever the administrative programme looks like.
How should we measure first-year retention? By tenure band rather than as a single organisational rate. First-year attrition has different causes and different costs from attrition at year four, and averaging them hides both.
Does this apply to internal moves as well as new hires? The meta-analysis distinguishes school-to-work from work-to-work transitions as a methodological moderator, so the literature treats them as potentially different. An internal mover starts with social acceptance and without role clarity, which suggests the effort belongs in a different place, but this evidence does not settle it.
Sources
- Rubenstein, Eberly, Lee & Mitchell (2018), Surveying the forest: a meta-analysis, moderator investigation, and future-oriented discussion of the antecedents of voluntary employee turnover, Personnel Psychology 71(1), 23-65
- Bauer, Bodner, Erdogan, Truxillo & Tucker (2007), Newcomer adjustment during organizational socialization: A meta-analytic review of antecedents, outcomes, and methods, Journal of Applied Psychology 92(3), 707-721
- Portland State University repository record for the newcomer adjustment meta-analysis, including the abstract
- US Bureau of Labor Statistics, Employee Tenure Summary, January 2024
