Attrition

Attrition, employee turnover and churn are used interchangeably and mean the same thing: people leaving. The distinctions worth maintaining are different ones.

An organisation that reports a single attrition figure without these distinctions cannot act on it, because the response to each is different. High involuntary attrition points at selection. High early attrition points at the promise made at hiring. High regretted attrition among long-tenured staff points at progression or at pay compression.

The visible costs are recruitment fees, advertising and the time spent interviewing. They are the smaller part.

• The vacancy period. Work is redistributed to colleagues, who are then more likely to leave themselves, or it does not get done.
• The productivity ramp. A replacement takes months to reach the output of the person who left, and longer in specialist or relationship-heavy roles.
• Knowledge that was never written down. This is the largest cost and the hardest to quantify, and it is concentrated in exactly the long-tenured people organisations most regret losing.
• Relationships. In client-facing roles the departing person may take the relationship, and occasionally the client.
• The signal to others. One respected person leaving prompts others to look, which is why attrition clusters in time as well as by team.

The common rule of thumb, that replacing someone costs some multiple of their salary, is a serviceable prompt and a poor estimate. What it gets right is that the cost is far larger than the recruitment invoice, which is the only part most organisations actually see.

Exit data across organisations converges on a short list, and the order is fairly stable.

• The manager. The most consistent single factor in voluntary attrition. People leave situations they cannot influence, and the manager controls most of what an employee experiences daily.
• Progression. Not promotion specifically, but the absence of a visible path. This shows up at the two to three year mark.
• Pay, usually as compression rather than absolute level. Someone who discovers a new joiner at their grade is paid more does not negotiate; they look.
• Workload without control. Long hours are tolerated where people can influence what they work on and are not where they cannot.
• Role clarity. Employees who cannot say what success looks like disengage before they resign.
• The promise made at hiring. Where the job differs materially from how it was described, the exit is usually early and the cause is recruitment.

Pay is cited more often than it causes, because it is the easiest thing to say in an exit interview and the hardest for the employer to argue with. When exit data says pay and the pay is competitive, the real cause is usually one of the others.

Most retention initiatives target the wrong moment. By the time someone resigns the decision was made months earlier, and a counter-offer at that point buys a few months and rarely more.

• Act on the manager dimension. Attrition concentrated under particular managers is the single most addressable pattern, and it is usually visible in data the organisation already holds.
• Fix compression deliberately. Review existing staff against what new hires are being offered at the same grade, before someone discovers the gap.
• Make progression visible even where promotions are scarce. People tolerate slow progression far better than invisible progression.
• Use the first ninety days properly. Early attrition is the cheapest to prevent and the most often ignored.
• Ask people who stay, not only people who leave. The interviews worth having are with the people who considered leaving and did not, and almost nobody conducts them.

Finally, accept a floor. Some attrition is healthy: it creates movement, brings in new capability and allows people to leave roles they have outgrown. The goal is not zero, it is that the people leaving are not disproportionately the ones you wanted to keep.