Employee Retention Guide

Because a large share of avoidable first-year departures were decided during hiring rather than during employment. Three causes recur. The role was oversold, so the person arrived expecting something the job does not contain. The job itself was designed badly before anybody was hired into it, with incoherent scope, no authority to match the responsibility, or a workload two people previously shared. And the profile hired was wrong for the actual work, usually because the scorecard described an idealised version of the role rather than the real one. None of those are fixable by engagement programmes afterwards. They are fixable at intake, in the job description, and in what gets promised at offer stage. The economics are straightforward too: with average cost per hire at $5,475 according to the SHRM 2025 Benchmarking Report, somebody leaving in month seven means paying that again, plus the vacancy and the ramp. Model your own figure with an employee turnover calculator using your own inputs rather than a borrowed multiplier. Start with any role that has turned over twice, because a repeat departure from one seat is a design problem.

As a resignation somewhere between month three and month nine, usually explained as a better opportunity elsewhere. What actually happened is a gap between what was described and what arrived. The defence is a realistic preview given deliberately: say the hard parts out loud during interviews, describe the least enjoyable part of a normal week, and let the candidate speak to somebody doing the job today without their manager in the room. Anyone who withdraws after that conversation was going to leave anyway, and losing them before an offer costs a fraction of losing them after a relocation. Then write down what was promised. Scope, title path, remote policy, on-call, travel, review cycle, and anything said about progression. Hand that document to the manager on day one, because most broken promises are not dishonesty; they are a conversation the manager never heard about. The role-specific hiring guides are a useful check on whether the role you are describing matches how that job is normally structured elsewhere. Then check the promise against reality at ninety days, while there is still time to close the gap or explain it.

It decides how quickly somebody becomes useful and whether they build the relationships that make leaving feel costly. The mechanics are unglamorous and they compound. Access, equipment and accounts working on day one, arranged the week before rather than the morning of. A named buddy who is not the manager. A manager one-to-one inside the first three days. A written 30, 60 and 90 day plan with one real deliverable in the first month, because early contribution is what turns a new hire from a guest into a member of the team. Introductions arranged rather than left to chance, which matters most for remote joiners with no corridor to bump into anyone in. Then two honest check-ins, at 30 and 90 days, asking what differs from what they expected. Most first-week failures start with a poor handover from the applicant tracking system to whoever runs onboarding, so make that handover a defined step with an owner rather than an email. Measure completion of the plan rather than attendance at an orientation session, since the plan is the part that produces a contributing colleague.

Timing, and therefore usefulness. A stay interview happens while the person is still there and something can still change. An exit interview happens after the decision, filtered by politeness, a reference they may need later, and a wish to leave cleanly. Exit interviews are still worth running, but read them in aggregate for patterns rather than treating any single one as the truth. Stay interviews are the higher-value practice and hardly anyone runs them. Twice a year, thirty minutes, ideally with somebody other than the direct manager where trust is the issue. The rule that makes them work: act visibly on at least one thing raised, within a month. A stay interview that produces no change teaches people that speaking up is pointless, which leaves you worse off than never asking. Keep notes out of the performance record so the conversation stays safe to have honestly. Run them at a consistent point in the year rather than in response to a resignation, because a stay interview scheduled the week after somebody quits reads as damage control to the whole team.

• What would make you seriously consider leaving?
• What part of the role would you change tomorrow if you could?
• What do you want to be doing in two years, and is it possible here?
• Has anything promised when you joined not materialised?
• What is the best thing about this job that I should protect?

A great deal, because the manager controls nearly everything experienced daily: workload, feedback, which projects land where, whether credit is distributed fairly, and whether growth conversations happen at all. Rather than argue about the size of the effect in general, measure it inside your own company. Look at voluntary attrition by manager across two years, adjusting for team size and function, and look at internal transfer requests out of each team. Patterns appear quickly and are rarely a surprise to anyone working there. The most common root cause is structural: promoting the strongest individual contributor into management with no training and no support, then evaluating them on output rather than on team health. Fix it in three places. Make management a deliberate choice with a genuine alternative track. Train the specific behaviours, which are one-to-ones, feedback, workload distribution and career conversations. And be willing to move somebody out of management without framing it as a demotion, which is the part most companies avoid. Give new managers a peer group as well as training, since most management questions are too small to escalate and too specific to look up.

Because most people who leave for a new challenge would have taken that challenge internally if it had been visible and reachable. Three mechanics make it work. Post roles internally before or alongside external advertising, and mean it rather than running a formality after the decision is made. Set a transfer policy that stops a manager blocking a move indefinitely, with a maximum handover notice, because talent hoarding is the single most common reason these programmes fail. And keep a record of skills and interests, so a recruiter filling a role thinks of internal people first. Career pathing supports all of it: two tracks, individual contributor and management, with the expectations at each level written down, and lateral moves treated as normal rather than as a step sideways. Internal hires arrive with context, relationships and trust already in place. Note that this data usually lives in the HR system rather than the recruiting one, so read the ATS versus HRIS distinction before assuming a single tool covers both. Count internal fills as a hiring outcome in the reporting, because whatever is not counted stops being anybody's priority.

On a fixed calendar, at least annually, using market data you actually purchase rather than whatever a candidate claims in a negotiation. The failure mode is compression: pay for new hires drifts up with the market while existing people stay on their joining number, until somebody internal discovers what a new colleague earns. That discovery produces resignations that look sudden and were not. Review bands, check for compression explicitly, and correct it during the review rather than when a resignation forces it. Counter-offers deserve a plain position: they usually fail as a retention strategy, because money is rarely the only reason, and paying somebody who resigned teaches everybody else the mechanism for a raise. Some companies still make them for critical roles, and if you do, set the policy in advance rather than deciding in the moment. Pay transparency and pay equity obligations vary by jurisdiction and change often, so confirm what applies to you with qualified employment counsel. Explain the outcome of the review even when nothing changes for a given person, since silence after a stated review cycle gets read as a decision anyway.

One turnover percentage tells leadership almost nothing. Split it four ways. Voluntary versus involuntary, because those have different causes and different fixes. Regretted versus non-regretted, decided by asking the manager a single question before the person leaves: would you rehire them. By tenure cohort, since first-year attrition is a hiring and onboarding signal while fifth-year attrition is a growth and pay signal. And by manager, team, source and role, which is where the actionable patterns live. Then price it. Turnover cost is replacement cost plus the cost of the vacancy plus the ramp before a replacement is fully productive, so model it with a cost of vacancy calculator using your own salary and revenue inputs. Resist grading yourself against a published benchmark; healthy turnover varies enormously by industry, seniority and region. Compare against your own trend and your own cohorts, and report retention beside your other recruitment metrics rather than in isolation. Report it alongside a short note on what changed, because leadership seeing attrition numbers with no action attached stops asking within two quarters. Keep the definitions fixed so the trend stays comparable.