Fill in these first
- Place and what is served: [e.g. a diagnostic lab with three sample collection counters]
- Queue log: [paste rows: token number, service type, time issued, time called, time service ended, counter]
- Days covered: [e.g. two full weeks]
- Opening hours and counters open by hour: [paste]
- Wait time target: [e.g. 90 percent of people called within 15 minutes]
Task
Show where and when waits go wrong and what staffing would fix them.
Steps
- Wait time is time called minus time issued; service time is time ended minus time called.
- Median and 90th percentile of each, by hour of day, day of week, service type and counter.
- Arrivals per hour against people served per hour to find when the queue grows.
- Counters needed each hour so service capacity covers arrivals with the target met.
Output format
- | Hour | Arrivals | Median wait | 90th percentile wait | Counters open | Counters needed |
- A day by hour heat table of the 90th percentile wait.
- Service types or counters that take much longer than the rest, with possible reasons.
- Five changes ranked by effect: staffing by hour, a separate counter for quick services, appointments, pre-filled forms, or others the data supports.
Rules
- Exclude tokens never called from wait times and report them as walk-aways.
- Show the method used for counters needed in two or three lines.






