Most resignations are about the immediate manager

Turnover is discussed as a market problem and is largely a local one. The variation between teams inside a single organisation is usually greater than the variation between employers.

When an organisation loses people it tends to reach for external explanations: the market is competitive, salaries have moved, this generation changes jobs more often. Some of that is true, and it does not explain the thing that is visible in almost every organisation's own data, which is that turnover is not distributed evenly. Some teams lose nobody for years. Others lose several people annually while doing similar work, for the same pay, under the same policies, in the same building.

That variation is the most useful information an organisation has about why people leave, and it is generally not examined, because examining it means arriving at a conclusion about a named manager. It is considerably more comfortable to conclude that the labour market is difficult, and the comfortable conclusion also has the advantage of implying that nothing internal needs to change.

The mechanism is not usually dramatic. Managers who lose people are rarely tyrannical; that variety is uncommon and gets dealt with. The far more common version is a manager who is absent, who does not give feedback until the annual review, who takes credit or fails to distribute it, who cannot make a decision, or who is simply too busy to notice that somebody is struggling. None of that produces a complaint, and all of it produces a steady loss of people who could not name a specific grievance.

The second thing the data usually shows is when people leave, which clusters. There is a spike in the first few months, driven almost entirely by the gap between what the role was described as and what it turned out to be, and by whether anybody organised the person's arrival. There is another around the point at which somebody concludes they will not progress here, which occurs at a fairly predictable interval depending on the field. Both of those are addressable and both are cheaper to address than to absorb.

The third is that pay matters in a specific and limited way. It is very difficult to retain somebody who is being paid materially below what they could get elsewhere, and beyond that threshold additional pay buys surprisingly little retention. What it does buy is a delay, and a person who has decided to leave for other reasons and accepts a counter-offer typically leaves within the following year anyway, which is well enough documented that counter-offers are a poor use of money.

This does not make retention easy, and it should not be read as implying that people leave for soft reasons. They leave because of workload that never lets up, because promises were not kept, because they watched somebody less capable be promoted, because their manager did not defend them, or because their circumstances changed and the job could not accommodate it. Those are substantial reasons and most of them are within an organisation's control.

The practical position is that retention work is mostly management work rather than human resources work, and that the highest-yield interventions are unglamorous: selecting and supporting managers properly, running a decent first month, telling people the truth about progression, and noticing when somebody's circumstances have changed. None of that is a programme, which is part of why organisations reach for programmes instead.

One further point about why the comfortable explanation persists. Concluding that the labour market is difficult requires nothing of anybody and can be said in a board paper without consequence. Concluding that three teams account for most of the departures requires somebody to have a conversation with three managers who are otherwise performing well and who will not accept the finding. The second is correct considerably more often and it is nobody's preferred task, which is why organisations that examine turnover honestly are the exception rather than the norm.