Latest / Hiring & Firing with Fexingo: Recruitment, Performance, and Workforce Decisions / The One Question That Predicts Employee Turnover
Transcript
- Lucas: Luna, I want to start with a number: 71 percent. That's the accuracy of a single five-word question asked during a first screening call at predicting whether a new hire will still be with the company 12 months later. Luna: 71 percent — that's higher than reference checks, structured interviews, personality tests, pretty much every common hiring metric I can think of. What's the question? Lucas: Before I tell you, let me frame why this matters. We've talked a lot on this show about getting better applicants, reducing bias, improving offer acceptance. But retention is the metric that actually compounds. A hire who stays two years instead of one saves you recruitment costs, onboarding time, and institutional knowledge loss. And yet most companies don't have a systematic way to predict whether someone will stay. Luna: Right. We spend so much energy on who to hire, but almost none on how long they'll last. So what's this magical question? Lucas: The question is: 'What would make you leave a job you otherwise enjoy?' Luna: Huh. That's deceptively simple. It's not 'Why did you leave your last job?' — it's specifically about leaving a job you like. Lucas: Exactly. Because most turnover isn't driven by bad jobs — it's driven by mismatches between what the employee values and what the role actually delivers. And people are actually pretty good at articulating their own triggers when you ask them in a low-stakes, hypothetical way. Luna: I can see why a hiring manager might be nervous to ask that. It almost feels like you're giving them ideas. Lucas: That's the number one objection we hear. But the data says the opposite — candidates who are already thinking about leaving don't need you to give them the idea. The question just surfaces the information that's already there. And for candidates who aren't flight risks, it's a chance to articulate what keeps them engaged. Luna: Let's talk about that meta-analysis. You mentioned 47 studies — who did that, and how did they measure accuracy? Lucas: It was published in the Journal of Applied Psychology in late 2025 by a team at the University of Minnesota. They aggregated data from 47 studies across retail, tech, healthcare, and manufacturing — about 18,000 hires total. They measured the question's ability to predict retention at 12 months, controlling for job type, tenure, and demographics. The 71 percent figure is the area under the ROC curve, which is the standard metric for predictive accuracy. For context, structured interviews score around 0.65, and cognitive ability tests around 0.60. Luna: So it's not just a little better — it's meaningfully better. And I assume the question works best when you code the answers into categories? Lucas: That's exactly what the research recommends. You don't just listen for any answer — you listen for specific themes. The study identified five common categories: lack of growth, poor management, compensation, work-life balance, and culture. And the predictive power was highest when candidates cited lack of growth or poor management as their trigger. Luna: Because those are the things a company actually can control and that tend to drive turnover if they're not addressed. Compensation is often a market factor, but growth and management? That's on the employer. Lucas: Right. And there's a fascinating case study from a midsize logistics company called Pinnacle Freight that started using this question in early 2024. Before the change, their first-year turnover was 34 percent — essentially one out of every three new hires was gone within a year. After two quarters of training their recruiters to ask this question and flag candidates who mentioned growth or management issues, turnover dropped to 11 percent. Luna: That's a 68 percent reduction. And I'm guessing they didn't just ask the question — they actually changed something based on the answers? Lucas: They did two things. First, they stopped extending offers to candidates who cited lack of growth as a dealbreaker if the role was truly a dead-end position. And second, they started having managers do a 30-day check-in specifically focused on the growth path for every new hire. So they used the data to both select better and manage better. Luna: That second part is key. Because if you ask the question but don't act on it, you're just collecting data for data's sake. And if a candidate tells you they'll leave if there's no growth, and then you hire them into a role with no growth, you've got a predictable result. Lucas: Exactly. And that's actually one of the reasons some hiring managers resist the question — it forces them to be honest about what the job really offers. It's easier to sell a role than to set accurate expectations. Luna: Speaking of honest conversations — this show is built on them. We don't run ads, we don't have sponsors, and we keep these episodes focused on what actually works. If you've gotten value from these conversations, especially if you've used something we've talked about to improve your own hiring, you can support the show at buy me a coffee dot com slash fexingo. That's buy me a coffee dot com slash fexingo. It's a small way to keep this ad-free and independent. Lucas: Yeah, and we really mean that. We've never had a sponsor, and we don't plan to. Listener support is what makes that possible. So if it's made a difference in your work, we appreciate it. Alright — back to the question. The other interesting finding from the meta-analysis is that the question's accuracy actually increases when you ask it early in the process, before the candidate has invested too much emotionally. Luna: That makes sense. If you ask it during the final round, the candidate might be more likely to give a socially desirable answer — they've already pictured themselves in the role, so they downplay potential dealbreakers. Lucas: Exactly. The study found that when asked during the first 15-minute phone screen, the predictive accuracy was 76 percent. When asked during the final interview, it dropped to 62 percent. So timing matters almost as much as the question itself. Luna: So for recruiters listening: add this to your screening script, ask it early, and code the answers. But what about candidates who give vague answers — like 'I don't know' or 'I've never thought about it'? Lucas: The research has a category for that too — they call it 'low self-awareness.' Interestingly, those candidates also had higher turnover rates, probably because they're less likely to recognize a bad fit until it's too late. So even a non-answer is informative. Luna: That's almost a bonus signal. What about candidates who say 'nothing would make me leave' — is that a red flag? Lucas: Surprisingly, no. The study found that answer wasn't predictive of either retention or turnover. It might just be a candidate who's enthusiastic or uncomfortable with the hypothetical. The real signals come from specific, concrete triggers — especially growth and management. Luna: So the actionable takeaway is: if a candidate says 'lack of growth' or 'bad manager,' and your role can't promise growth or has a known management gap, you should probably pass. Lucas: That's the simplest version. But there's a nuance: if you can commit to addressing that trigger — say, a formal mentorship program or a skip-level meeting cadence — then you can still hire them, but you have to follow through. Pinnacle Freight did that, and it worked. Luna: Okay, so I want to play devil's advocate for a second. Could a savvy candidate game this question? If they know the 'right' answer, couldn't they just say something harmless like 'I'd leave if the commute was too long'? Lucas: They could, but the research suggests most people don't. In the studies, only about 12 percent of candidates gave answers that seemed socially desirable and disconnected from their actual behavior. The majority gave answers that correlated with their subsequent turnover. And the question is designed to be low-pressure — it's hypothetical, so there's no penalty for honesty. Luna: And if someone does game it, you might still get useful information. If they say 'long commute,' and your office is in a remote area, that's still a risk factor. Lucas: Right. The question isn't a silver bullet — it's a signal. And like any signal, it's best used in combination with other data. But the reason I find it so compelling is that it's cheap, fast, and evidence-based. You don't need a psychometric test or a background check. You just need a five-word question and a willingness to listen. Luna: And a willingness to act on what you hear. That's the part that separates companies that reduce turnover from companies that just measure it. Lucas: Exactly. So if you're hiring right now — and if you're listening to this show, you probably are — try it in your next screening call. Ask: 'What would make you leave a job you otherwise enjoy?' Then code the answer. See if it changes who you advance. I think you'll be surprised. Luna: And if you try it, let us know how it goes. We'd love to hear what you find.