Most hiring processes are designed to identify capability, not longevity. These five signals have disproportionate predictive value for retention — and most of them show up before the first interview.
The most reliable retention predictor is a genuine warm referral.
This isn't a heuristic — it's in the data. Referred employees consistently stay longer than those hired through job boards — research across multiple studies finds a meaningful and repeatable retention advantage. The reason isn't primarily about candidate quality, though that matters. It's about the nature of the endorsement.
When someone with a professional reputation endorses a candidate, they're making a claim they'll be held to. They know the company will trace the result back to them. That accountability filters the candidates they'll vouch for. They don't refer people they're uncertain about — the social and financial cost is too high.
The further implication: the more specific the introduction, the more weight it carries. A referrer who says "we built the Series A data architecture together and she was the one who delivered it" is making a different claim than someone who says "I know her and think she'd be great."
Candidates who can name one concrete reason why this particular role, at this particular company, at this particular stage is interesting to them tend to stay longer.
The reason doesn't have to be elaborate. But it should be specific. "I want to build enterprise B2B experience" or "I've been following your product since the 2023 redesign and I think the problem you're solving is underserved" is a different answer than "I'm looking for my next challenge."
Specificity signals that the candidate has thought about whether this is the right move — not just whether they want a job. Candidates who land at a company because it was the best offer available at the time tend to move again when the next offer comes along. Candidates who chose deliberately tend to stay until the reason they chose stops being true.
Patterns in employment history predict future patterns. A candidate with a consistent 3–5 year tenure at previous companies is more likely to stay 3–5 years than one with six jobs in six years.
This isn't a rule — there are legitimate reasons for shorter stints (company shutdowns, acquisitions, layoffs, deliberate career pivots). The question to ask is whether the tenure pattern is consistent with what you know about the candidate's career narrative.
What to look for: one or two long tenures interrupted by legitimate short ones, or a clear narrative arc that explains the moves. What's less predictive: a pattern of leaving after 10–14 months across multiple companies, which typically reflects a different relationship to employment stability.
Warm referrals vary enormously in quality depending on how well the referrer actually knows the candidate.
A hiring manager who sees an introduction from someone who "worked at the same company" in a different department has very different signal than one from someone who "co-led the migration that shipped under budget in Q3."
The best indicator of retention isn't that the candidate is known — it's that they're known specifically for doing work relevant to the role, in conditions similar to what they'll face.
This is why the written introduction matters more than most hiring teams realise. A specific, well-written referral describes work context. An unspecific one doesn't. The former predicts retention; the latter predicts capability without context.
During interviews, candidates who ask substantive questions about the team they'll join, how the team makes decisions, and what the working culture actually looks like tend to be evaluating fit — not just offers.
Candidates primarily interested in compensation maximisation will typically accept a better offer if one arrives. Candidates interested in team and environment will factor in the cost of disrupting relationships they've invested in.
This is not an argument against paying well — you should pay well. It's an observation that candidates who ask compensation-only questions during early screening are signalling something different than candidates who ask about team size, collaboration style, and how priorities get set.
The referral process gives you a preview of this. A referrer who says "she always asks about the team before she asks about the package" is telling you something predictively valuable.
Of these five signals, the first — a genuine, specific warm referral from someone with professional accountability — is the one that requires the least judgment and has the most consistent predictive value.