Turnover is the proportion of staff who leave over a period, usually a year.
Calculating it
Departures during the period, divided by average headcount. Six people leaving from an average of twenty is thirty per cent.
In a small business a single departure moves the figure a great deal, so the trend across years is the only reading worth taking.
Distinguishing the kinds
Voluntary — they chose to go. This is the number that says something about the business.
Involuntary — they were dismissed. High involuntary turnover usually points at hiring or onboarding rather than at the people.
Regretted — whether the business wanted to keep them. Losing people it wanted to keep is a different problem from losing people it did not, and an undifferentiated rate hides which is happening.
Early turnover is its own signal
People leaving within the first few months point at the hire or the start rather than the job — see employee onboarding.
That is the cheapest turnover to fix and the most commonly ignored, because each instance looks like an individual mistake.
What it costs
Recruiting, the vacancy, and the ramp while a new person produces less than full output. In skilled trades that ramp is months.
Against that, the pay rise or the schedule change that would have kept someone is usually small — see employee retention.
Absence
Rising absence frequently precedes departures and is worth watching for the same reason: it is an early version of the same signal.
Exit conversations
Too late to keep that person and still worth having. The pattern across several is more reliable than any one of them, and people are more candid about the job than about their manager.
The stronger source of the same information is asking current staff, which almost no small business does.