THE USEFUL ANSWER
Coverage says someone is assigned. Capacity asks whether the assigned people can handle the arriving work. Plan both, using the busiest meaningful interval rather than the weekly average.
- Estimate work in minutes, not just message counts.
- Reserve time for breaks, review and handover explicitly.
- Use observed queues to adjust assumptions; this arithmetic is not a waiting-time forecast.
Count the work you actually handle
A message is not always a unit of work. One short question may take seconds; another thread may require reading history, checking an approved offer and asking a supervisor. Counting both as one message hides the difference.
Choose a unit that your team can observe consistently, such as a conversation task completed during a defined interval. Include reading, composing, review and necessary record updates in handling time. Explain whether multiple replies in one conversation count as one task or several.
Use the coverage planner for staffing hours, then add an interval-level worksheet for workload. The two calculations answer different questions and should remain visible together.
Work through one peak hour
Suppose an invented inbox receives 24 tasks in an hour and each requires four minutes of operator work. Demand is 24 × 4 = 96 minutes. Two assigned operators provide 120 scheduled minutes, but assume only 75% of that time is available for direct inbox work. Available capacity is 120 × 0.75 = 90 minutes.
The difference is six work minutes. Under these simplified assumptions, unfinished work grows during that hour even though two names appear on the rota. This example is arithmetic, not a measured agency benchmark or a prediction of exact subscriber waiting times.
| Input | Example value | What to validate locally |
|---|---|---|
| Arriving tasks | 24 per hour | The definition and actual arrival pattern |
| Handling time | 4 minutes per task | Whether review and reading are included |
| Assigned people | 2 | Whether both are dedicated to this inbox |
| Direct-work availability | 75% | Breaks, meetings and other duties |
| Required work | 96 minutes | Sensitivity to a busier interval |
Do not subtract review time twice: either include it in handling time or account for it in availability, with the distinction documented.
Avoid the weekly-average trap
A week can have enough total staffing hours and still have the wrong people scheduled at the wrong time. A quiet morning does not retroactively answer a queue that formed the previous evening.
Look at a useful interval size, such as an hour, and shorten it if short bursts matter to the operation. Record arrival volume, active staff and unfinished work at the start and end. A growing queue is a signal to investigate, not automatic proof that a person is working too slowly.
Separate new tasks from work carried into the interval. Otherwise the same queue can appear as mysterious excess demand every time a shift changes.
Reserve flexibility without inventing precision
The simple model assumes average handling time and usable staffing minutes. It does not model random arrivals, task dependencies, interruptions or the distribution of long conversations. A ratio below 100% therefore does not guarantee a particular response time.
Use a low, expected and busy scenario. If handling time rises from four to five minutes in the example, the same 24 tasks require 120 minutes. That is 30 minutes above the assumed 90-minute capacity. The sensitivity matters more than reporting the first estimate to two decimal places.
A practical response might be moving a break, narrowing nonessential tasks or adding reviewed support during the peak. Validate the change with observed work and response outcomes.
Connect the calculation to the rota
Assign an owner to each interval and specify who can help when the queue exceeds the agreed threshold. Define the threshold in terms your team can see, such as oldest unresolved task and count of tasks awaiting action.
Time zones can move the apparent peak between teams; the time-zone planning guide covers that boundary. The handover audit covers what happens to unfinished work when one shift ends.
Keep the first worksheet small. A clear set of assumptions and one week of consistent observations is more actionable than a complicated capacity number whose inputs nobody can explain.
Sources & editorial notes
Primary references checked on 10 September 2026. Calculations and proposed workflows are our editorial examples, not independently observed provider results.