Voluntary turnover is one of the most expensive recurring costs in scaling businesses and one of the most systematically misdiagnosed. The research on what actually drives departure, and why the standard methods for identifying it are structurally flawed, produces a significantly different picture from what most attrition conversations assume.

Why salary increases don’t produce retention

Herzberg, Mausner and Snyderman’s (1959) two-factor theory established a distinction that the attrition intervention literature has not fully absorbed. The factors that produce job dissatisfaction — hygiene factors, including salary, working conditions, company policy, and supervision quality at its most basic level — are different in kind from the factors that produce genuine motivation and retention. These are not opposite ends of the same scale; they are different variables. Removing dissatisfaction does not produce satisfaction. It produces neutral.

The practical implication for attrition is direct: salary increases, benefits improvements, and working condition enhancements address the hygiene factors that produce dissatisfaction when absent. They are necessary but not sufficient for retention. The team member who is leaving because they have no autonomy, no development, and no meaningful feedback will not be retained by a salary increase. The salary increase removes one reason to leave; it does not address the reasons they are actually leaving.

The manager relationship and what the Gallup data established

Harter, Schmidt and Hayes’s (2002) analysis of 7,939 business units and 198,514 employees found that employee engagement — the strongest available predictor of voluntary turnover — was substantially determined by the quality of the direct manager relationship. The variance in engagement within companies, between different teams led by different managers, exceeded the variance in engagement between companies. The team’s experience is more a function of who leads it than of which company they are in.

The research from this series converges on the same conclusion from multiple directions. The Pygmalion and Golem effect research establishes that the manager’s expectations shape performance through four specific behavioural channels. The psychological safety research establishes that the manager’s vulnerability modelling determines whether the team can share information honestly. The conflict avoidance research establishes that unaddressed interpersonal problems accumulate compound interest. The control research establishes that the autonomy the team experiences is primarily a function of the manager’s delegation behaviour.

The team member who leaves is typically leaving the experience that the manager’s behaviour has created — the expectation climate, the safety level, the feedback quality, the development investment, the degree to which their autonomy is respected — not the company’s strategy, culture, or compensation structure. The departure is attributed to compensation in the exit interview because that is the explanation that requires the least disclosure of the actual experience.

Why exit interviews are structurally unable to identify the actual causes

Giacalone, Elig and Knouse’s (1997) research on exit interview validity established the two mechanisms that produce systematic underreporting. The departing employee has reputational and reference concerns: honest criticism of the manager or the leadership team is risky when the same people may be contacted for references. The rational threat assessment that Detert and Edmondson identified as producing silence in the team environment operates with equal force in the exit interview — the departing person will not say what they cannot safely say.

The Nisbett and Wilson (1977) post-hoc rationalisation research adds the second mechanism: even when the departing employee wants to be honest, the stated reason for leaving is a narrative constructed after the decision rather than an accurate report of the actual causal mechanism. “I found a better opportunity” is always partially true and rarely the complete picture. The better opportunity was pursued because something in the current situation was pushing — but the push is harder to articulate, more personal, and more risky to disclose than the pull.

The exit interview that asks “why are you leaving?” is structured to receive the answer the departing person is willing to give in the context of a conversation with someone who works for the company they are leaving. The true answer is typically in the data that precedes the departure.

The leading indicators that exit interviews miss

The Maslach burnout dimensions — emotional exhaustion, cynicism, and reduced efficacy — are the psychological states that typically precede voluntary departure by months. They are measurable in current employees before the departure decision is made, and they are substantially caused by the manager-relationship variables that the exit interview captures too late and too inaccurately.

Ryan and Deci’s SDT research predicts the specific needs whose chronic unmet state produces these burnout dimensions and eventually departure: autonomy, competence, and relatedness. The team member who is micromanaged has unmet autonomy need. The team member who is not developing has unmet competence need. The team member in a conflict-avoidant, low-safety team has unmet relatedness need. These are the actual mechanisms of departure — they are measurable before the departure decision, and they are addressable by the manager whose behaviour is producing them.

The Gallup Q12 engagement items are the validated operational tools for this measurement: “my manager cares about me as a person,” “I have someone at work who encourages my development,” and “my opinions count” are the items most directly predictive of departure. They measure the experience that the research identifies as the actual attrition driver, while there is still time to change it.

The stay interview — a structured conversation with current employees about what would make them leave and what keeps them — produces more accurate data than the exit interview because it lacks the reference concern that suppresses honesty at departure. The departing employee cannot say what they know; the current employee who trusts the environment can.

Books worth reading on this

An Elegant Puzzle by Will Larson. Larson’s account of engineering management in scaling companies — including the specific management practices that retain and develop capable technical people — is the most practically structured available treatment of what the research-supported retention behaviours look like in daily practice. His specific account of the systems and rhythms that effective managers build to ensure development, recognition, and genuine feedback provides the most actionable complement to the research framework this article describes. It’s the Manager by Jim Clifton and Jim Harter. The most recent synthesis of Gallup’s research on the manager relationship and its consequences for retention, engagement, and performance — written specifically for the organisational context where most attrition prevention decisions are made. Clifton and Harter’s specific account of the twelve items that predict team outcomes, and the manager behaviours that determine each of them, is the most directly applicable available operational guide to the attrition mechanisms this article describes.

If the dynamics described here are significantly affecting your wellbeing, speaking with a psychologist is the right next step. UK: Samaritans (116 123, free, 24/7). Mind (0300 123 3393). BACP: bacp.co.uk/search/Therapists. Crisis Text Line — text HOME to 741741 (US, UK, Canada, Ireland). International: internationaltherapistdirectory.com.

This article is for educational and informational purposes only. Sources: Herzberg, F., Mausner, B. & Snyderman, B.B. (1959), The Motivation to Work, Wiley. Harter, J.K., Schmidt, F.L. & Hayes, T.L. (2002), Business-Unit-Level Relationship Between Employee Satisfaction, Employee Engagement, and Business Outcomes, Journal of Applied Psychology, 87(2), 268–279. Giacalone, R.A., Elig, T.W. & Knouse, S.B. (1997), Exit Surveys, in Applied Impression Management, Sage. Nisbett, R.E. & Wilson, T.D. (1977), Telling More Than We Can Know, Psychological Review, 84(3), 231–259. Edmondson, A.C. (1999), Psychological Safety and Learning Behavior in Work Teams, Administrative Science Quarterly, 44(2), 350–383. Deci, E.L. & Ryan, R.M. (2000), The “What” and “Why” of Goal Pursuits, Psychological Inquiry, 11(4), 227–268. Rousseau, D.M. (1989), Psychological and Implied Contracts in Organizations, Employee Responsibilities and Rights Journal, 2(2), 121–139. Buckingham, M. & Coffman, C. (1999), First, Break All the Rules, Simon & Schuster. Larson, W. (2019), An Elegant Puzzle, Stripe Press. Clifton, J. & Harter, J. (2019), It’s the Manager, Gallup Press.