Imagine a service operation that needs 108 people-hours of work covered across nine one-hour intervals. Divide one by the other and you get a tidy answer: 12 people on average.

But the average has discarded something important: when the work arrives.

One day, two very different stories

Each value is the people needed throughout that one-hour interval.

Hour starting People needed Flat staffing
8am 6 12
9am 8 12
10am 12 12
11am 16 12
Noon 18 12
1pm 16 12
2pm 14 12
3pm 10 12
4pm 8 12

Both columns add to 108 people-hours. Yet at noon the flat plan is six people short. Across the whole day, it has 16 people-hours below demand and 16 above demand. Those do not automatically cancel in practice: spare capacity at 8am cannot serve a customer who arrives at noon.

The flat plan covers 92 of the 108 people-hours needed, or about 85%. Coverage here means the sum of the smaller of demand and staffing in each interval, divided by total demand.

Change the timing, then inspect the trade-off

The coverage experiment lets you vary a team’s size and the distribution of its availability. It uses the same demand curve, but a different staffing model from the flat baseline above: not everyone is available throughout all nine intervals.

Moving the timing slider redistributes a fixed total of available people-hours. Increasing team size adds capacity. Look at both the uncovered demand and the spare capacity; a single coverage percentage does not tell the whole story.

Bring the real-world constraints back in

This example leaves out skills, breaks, individual shift lengths, labour rules, productivity variation and uncertainty in the forecast. Real roster design needs those constraints. Where the task times behind them come from is its own question, and one slow observation can move a whole estimate. The purpose of the example is to expose one particular blind spot: a daily total can balance while the day itself does not.

Seeing the gap is the first half. Deciding what to do about it — more people, or the same people at different times — is the follow-up note.

Try the 30-second visual story, then consider this question with your team: at what interval does your current reporting stop being useful for a staffing decision?