What is on-time rate?
On-time rate is the share of a person's recorded arrivals that fell inside their allowed start window, as a percentage. Eighteen on-time arrivals out of twenty gives ninety per cent. It appears as an attendance headline figure beside the count of late arrivals that produced it.
The measure asks a narrower question than an attendance percentage. Not whether somebody came in. Whether they came in when expected.
Its level is almost never the interesting part. Somebody sitting at eighty-eight per cent all year is a person with a commute. Somebody sliding from ninety-five to sixty over two months is telling you something.
The figure is also sensitive to the width of the allowance it was measured against. Add five minutes of grace and everybody improves at once, without anybody changing their behaviour.
So the percentage only means something next to two other facts. The window that produced it, and how many arrivals it was built from.
The count of late arrivals sits beside it for that reason. A figure of eighty per cent over five days and the same figure over sixty days are different claims. One is a bad week. The other is a habit.
Sample size is where most misreadings start. A rate built on five mornings swings twenty points on one of them. Nobody should act on that, and plenty of people do.
A month is a reasonable floor before the number is worth discussing. Below that, read the raw arrivals instead. There are few enough to look at one by one.
Comparing the figure across teams on different shift patterns is rarely fair. A nine o'clock start in heavy traffic and a two o'clock start are not the same test, so the two percentages are not really comparable.
Read the figure against the same team's earlier periods rather than against another team. The month-over-month view is the comparison that survives scrutiny.
Related questions
How small a sample is too small?
A rate built on five arrivals moves twenty points on one bad morning. Wait for a month of records before reading anything into the number.
Should the figure be shared with the team?
That is a culture decision rather than a technical one, and publishing individual figures changes behaviour in ways worth deciding deliberately first.
How does it differ from an attendance percentage?
They count different failures. Somebody present on every expected day can still arrive outside the window on half of them.