Why People Leave

Organisations think they know why people leave because they ask on the way out. The exit interview is the least reliable instrument in the building.

Why exit interviews mislead

The person has nothing to gain from candour and something to lose.

References, future contact, an industry that is smaller than it looks. "A great opportunity came up" is the safe answer and it is the one most people give.

It is also conducted by the organisation being evaluated, frequently by somebody adjacent to the problem.

And it is a single retrospective account given at an emotional moment, which is not how anybody reconstructs a decision made over months.

What the data points to instead

Manager quality. Poor management is associated with a fourfold increase in turnover risk, and among people with ineffective managers 85% are reported as actively job seeking. (Vendor research.)

Career development, differing by stage. Early-career people want skill building; mid-career want visible trajectory; late-career want to remain relevant. The same conversation fails all three when it is generic.

Whether problems could be raised. Only one in three strongly agree they can speak up without fear, and people who cannot raise problems do not raise their intention to leave either.

And a change in conditions. Return-to-office mandates are associated with roughly 14% higher attrition, concentrated among those with options.

Where pay actually sits

Pay is the reason given and rarely the reason.

It is compared once, at offer, and then normalises. It becomes a live issue when something else is wrong and the person starts looking, at which point a competing offer supplies both the trigger and the explanation.

Which is why counter-offers have a poor record. They address the stated reason and leave the actual one, and the person leaves anyway within a year.

What to collect instead

Turnover grouped by manager, which requires no interview and is the single most informative view available.

Stay interviews — asking people who are not leaving what would make them consider it. Candour is higher because nothing is at stake.

Whether departures cluster after specific events. A policy change, a reorganisation, a promotion round.

And a consistent format, because the value is in comparison across time and the free-text exit interview cannot be compared with anything.

What to do with somebody who is already going

Ask anyway, and expect the polite answer.

Ask one specific question instead of an open one: what would have had to be different a year ago. The same discipline applies when reading a posting from the other side. It is narrower, it is easier to answer honestly, and it produces something actionable more often than "any feedback for us?"

The short version

For additional context on this topic, see Bloomberg.