The question, and the exit interview that doesn't explain it
The talent problem that most frustrates senior leaders is not the one they cannot solve. It is the one they keep solving wrong. Someone exceptional leaves. The exit interview is polite - compensation is mentioned, perhaps working conditions, sometimes a vague reference to opportunity. The real reason surfaces later: they joined an organization that will invest in who they can become. The previous one was managing them for who they already were.
Two bets placed on different timescales
Managing for performance is the act of maximizing what someone delivers in their current role, within their current capability, during the current period. It is a legitimate and necessary form of management. Its opposite is not development - it is tolerance for underperformance.
Investing in development is a categorically different act: a bet on what someone can become, taken before the evidence of that future capability is visible. It requires tolerating imperfect performance as capability is being built, and accepting that the return will arrive on a timescale that does not map onto the quarterly review cycle. The distinction is about which timescale the organization is managing: the capability it has today, or the one it will need in three years. Both timescales are real. Only one of them compounds.
Managing for performance optimizes for who someone is today.
Investing in development bets on who they can become.
The first extracts value. The second creates it.
Why the confusion is so durable - and so costly
Development is structurally harder to manage than performance. Performance produces visible, measurable outcomes within a defined period. Development produces ambiguous progress on a long horizon, passes through a phase of reduced output, and requires the manager to hold genuine uncertainty about the return. Most managers are evaluated on the performance of their teams in the short term. That incentive structure does not require anyone to be cynical for its effects to be systematic.
The people who suffer most from this confusion are precisely those with the highest potential - because they are the ones most sensitive to the distinction. They can tell whether their current role is developing them or merely deploying them. The second cost is less visible and more structurally dangerous: the people who stay. Those who remain in organizations that systematically prioritize performance over development tend to stabilize at the level of their current capability. The organization retains the person and loses the trajectory.
What genuine development actually requires
A diagnostic question, not a rating. Development asks: what is this person capable of becoming, what stands between them and that, and what does the organization need to build to close that distance? The organization that conflates them uses the performance review as a development tool - and produces neither effective performance management nor meaningful development.
Roles designed for stretch, not for confidence. Real development requires assignments that exceed current capability. Stretch roles are structurally different: the person operates at the edge of their competence, makes real mistakes in conditions with real stakes. That requires managers who can hold the discomfort of watching someone struggle.
Feedback given early enough to be useful. The most underused development instrument is the direct, specific, early observation shared with someone who is building capability - before the gap has become a performance problem.
The theory of the person
The leaders who build the strongest teams over time carry a genuine theory about each person's potential - what they are capable of becoming, what experiences they need to get there, what the organization owes them in return for the risk of stretching. The manager who can say to someone: "I think you are capable of X, and here is why I believe that, and here is what we need to build together to get there" - that manager is offering something that compensation cannot replicate.
The leader's real job: investor, not optimizer
An optimizer extracts maximum performance from existing assets. An investor allocates resources toward future capability - accepts real short-term costs in exchange for returns that are not yet visible. The result is a team that grows into something it is not yet. Most managers default to optimization, because its outputs are visible and its timeline matches the review cycle.
The bottom line
The best people are not primarily loyal to compensation. They are loyal to growth - to the concrete, ongoing experience of becoming more capable, more consequential, more themselves than they were the year before. The organizations that retain them are not necessarily the ones that pay most. They are the ones that make that experience credible.
The loss of someone exceptional to an organization that will actually develop them is not a compensation problem. It is a management problem - the problem of having managed someone entirely for what they were rather than for what they could become. The exit interview will not usually say this. The pattern, across years and people, eventually will.