Back to BlogIntelligence

Flight Risk: How to Spot Candidates Likely to Leave Within a Year

January 15, 20268 min read

The short answer

Flight risk is readable from a career history, but the useful version is specific rather than a single score. Look at the pattern of tenures rather than any one stint, whether their moves have been upward or lateral, whether the current employer is under visible stress, and whether their compensation sits below market for their level. Note the asymmetry: the same signals that make someone a flight risk for their current employer are what make them reachable for you, so the value depends entirely on which side of the desk you are on.

The most expensive hire is not the one who turns you down. It is the one who says yes, takes three months to ramp, delivers for another five, and then hands in their notice before the year is out. You paid the agency or the sourcing cost, you paid the salary, you paid the manager's time, and now you are back at the start of the same search. Early attrition is the quiet tax on hiring, and most of it is predictable.

There is a real tension here. The same restlessness that makes a candidate easy to recruit is the restlessness that makes them easy to recruit away from you nine months later. If you have read our companion piece on the signals that predict a yes, this is the other half of the same coin. Openness to move is an asset when you are closing and a liability once they are on your team. The job is to read both, not just the one that flatters your pipeline.

Signal 1: Short average tenure across the whole history

Read the pattern, not the single job. Everyone has one short stint they can explain, a startup that folded or a manager who left. What you are looking for is a habit. Three roles in a row of twelve to fifteen months each is a habit. That person is not unlucky, they are wired to move, and your role will not be the exception unless something in their life has changed.

Be fair about context. Early career people move faster, and that is normal. A 24 year old with two eighteen month stints is finding their feet. A 34 year old with five short stints is a pattern you should price in. And do not confuse contract or agency work with job hopping, because the tenure math looks the same but the story is completely different. This is exactly the kind of judgement that gets lost when you are skimming fifty profiles, which is why we push so hard on outcome based hiring rather than pattern matching on a resume alone.

Signal 2: Over-market pay and the ceiling problem

This one surprises people. A candidate you have to overpay to close is a candidate at higher risk of leaving, not lower. If someone is already sitting above their market band, the only way you win them is by paying even more, and now you have set a number that the market cannot easily beat except with a very aggressive counter. That feels like a win. It is actually a ceiling.

Here is how it plays out. You stretch to 55 LPA for a profile the market pays 40 for. Twelve months in they are bored or blocked, they go looking, and now they are anchored to 55 plus a jump, which means the next offer has to clear 70. Either they get it and leave, or they do not and they stew. Overpaying to close buys the yes and mortgages the retention. Pay fairly against the band and the loyalty tends to hold. If you are unsure what fair looks like, our breakdown of fair salary in Indian tech gives you the bands to anchor on.

The counter-offer trap

A candidate whose current employer is likely to counter hard is a flight risk in a specific way. They may accept your offer, use it to extract a raise, and never actually join. Even if they do join, a candidate motivated purely by money will do the same to you in a year. Watch for people at companies known to counter aggressively, and read whether the pull toward your role is about the work or only the number.

Signal 3: The recent promotion, which cuts the other way

Not every signal is a red flag. A recent promotion is a strong green one. Someone who was promoted in the last few months is sticky. The new title feels earned, the new scope is engaging, and the emotional reasons to leave have just been answered by their current employer. This is precisely why they are hard to recruit, and it is also why, once you do land a recently promoted person for the right reasons, they tend to stay.

The same logic applies to people who have recently taken on a visible project, a first management role, or a relocation their employer paid for. Life anchors reduce flight risk. A candidate who just bought a house near their office, or whose kids just started at a school in the same city, is not going to uproot on a whim. None of this shows up in a keyword search, which is the whole problem with keyword search.

Signal 4: The mismatch between ambition and the role on offer

The most avoidable early attrition comes from a role that is too small for the person. A senior candidate who takes an individual contributor seat because the money is good, when their whole history points toward leadership, will feel the mismatch within two quarters. They will do the work, they will do it well, and they will leave the moment a role with real scope appears.

  • Watch for a candidate whose trajectory is steeper than the role you are offering. They will outgrow it fast.
  • Watch for a domain jump the candidate is making for money rather than interest, because novelty wears off and the money stops feeling new.
  • Watch for a step down in company brand or scope that the candidate has not clearly reconciled. Prestige matters more than people admit.
  • Reward the candidate for whom your role is a genuine step up in scope, title or learning. That person has a reason to stay and build.

How TalentGPT scores retention and flight risk

Holding all four signals in your head across a full shortlist is not realistic, so in practice they get skipped and the flight risk shows up as a resignation letter. Inside TalentGPT every candidate carries a retention and flight-risk signal built from exactly these inputs: average tenure across their history, how their likely pay compares to their market band, promotion recency, and how well the role fits their trajectory. It runs across a pool of more than 300 million profiles, so you are reading the pattern on everyone, not just the three people you had time to check.

As with all our signals, the score is never a black box. Each candidate shows the reasoning in plain language, so you see that a low retention score comes from three short stints and a likely over-market ask, or that a high one rests on a recent promotion and a stable seven year run. You stay in control. The signal narrows a hundred names down to the ones worth a real conversation, and pairs naturally with the move-likelihood signal so you can find people who are both open to move now and likely to stay once they do. That combination is the quiet engine behind a shorter time to hire that does not come back to bite you two quarters later.

Frequently asked questions

Is a job hopper always a bad hire?

No, and treating short tenure as an automatic reject will cost you good people. The question is why the moves happened. Consecutive short stints chasing scope and learning read very differently from short stints chasing money or leaving under a cloud. Read the reasoning, not just the dates, and weigh whether your role gives them a reason to finally stay.

Why is over-market pay a flight risk instead of a retention tool?

Because it sets an anchor the candidate will want to beat next time. A person you overpaid to close is motivated by the number, and numbers are the easiest thing for a competitor to top. Fair pay against the band, combined with real scope, holds far better than a stretched offer that turns every performance review into a negotiation.

Can a signal really predict attrition before someone joins?

It predicts risk, not fate. No signal tells you a specific person will quit in month nine. What it does is separate the candidates whose history and situation point toward staying from those whose history points toward moving, so you can make a clear eyed decision and price the risk instead of being blindsided by it.

How do I balance flight risk against a candidate being easy to recruit?

Look at both signals side by side. The ideal hire scores high on move likelihood today and high on retention once they join, usually because your role is a genuine step up rather than a lateral cash grab. You can see both scores on the same profile in TalentGPT and rank your shortlist on the balance that fits the seat you are filling.

Frequently asked questions

How do you identify flight risk from a resume?

Read the pattern rather than the last job. A consistent history of moving every eighteen months is a pattern; one short stint among several long ones usually is not. Combine it with whether moves were upward or sideways, how long they have been in the current role relative to their own average, and whether the current employer is contracting.

Is job hopping always a bad sign?

No, and treating it as one filters out strong people. Early-career moves are often the fastest legitimate route to growth and pay, and some sectors move faster than others. What matters is whether each move looks like progression or escape, and whether the pattern is slowing as they become more senior.

What are the strongest predictors that someone will leave?

Being at or past their historical average tenure, a compensation gap versus market for their level, a stalled progression where peers moved up and they did not, and employer instability such as contraction or a funding gap. Any one is weak; together they are reasonably predictive.

How can I reduce flight risk in the people I hire?

Address the reason they left the last place, and make sure the role you are offering actually differs on that axis. If someone left for scope, more money will not retain them. Also be honest in the interview about what the job is, since most early attrition traces to a mismatch between the role described and the role that exists.

Should flight-risk signals be used to reject candidates?

Be careful. These are probabilistic patterns, not facts about a person, and rejecting on them alone will lose good candidates and can encode unfair assumptions. Use them to prioritise outreach and to shape the conversation, and let the interview rather than the pattern make the decision.