Plan capacity when demand rises and falls

Create a demand calendar

Plot the work across several recent operating cycles and the next planning period. Use meaningful units such as onboarding projects, support queues, store launches, or monthly closes. Mark predictable events, dependencies, and work that can be moved. Avoid averaging away the peak: an annual average can hide a two-week period when one specialist becomes the whole system’s constraint.

Find the constrained step

More people do not help if every item still waits for one approval, reviewer, or piece of equipment. Trace a typical unit of work from arrival to completion and note where queues form. The capacity plan should address that step first. Sometimes the answer is cross-training or clearer authority; sometimes a scarce capability genuinely needs additional coverage.

Choose the service promise

Decide what the team will protect when demand peaks: response time, quality review, high-priority customers, or a fixed deadline. Something else may need to wait. This is a leadership choice, not a scheduling detail. Without it, every request remains urgent and teams compensate through overtime or rushed decisions.

• What demand is stable enough to staff?
• Which peaks are predictable?
• Where does work actually queue?
• What can safely slow during a surge?

Match capacity to duration

Permanent hires support recurring work and retain context but create capacity after the peak if demand falls. Cross-training improves resilience while taking learning time. Fixed-duration support can cover defined surges but needs documentation and supervision. Process changes may reduce volume but rarely arrive instantly. Combine options rather than forcing every type of demand into one staffing model.

Test a peak plan

Illustrative example: an operations team has a steady onboarding flow and a predictable month-end surge. Mapping shows document review, not customer calls, causes the queue. The team cross-trains two coordinators for review, moves non-urgent reporting to the following week, and keeps one permanent specialist as owner. Hiring is reconsidered only if the elevated baseline continues across future cycles.

Define signals for changing the model

Agree what would trigger permanent hiring, additional cross-training, or a lower service promise. Use sustained workload and observed queues rather than one difficult day. Include quality and recovery time: a team that clears the peak only by postponing all improvement work is still carrying a cost. Review after each known surge, capture what moved, and update the next plan while the operational detail is fresh. Record the signal owner so evidence does not disappear between cycles. Keep the threshold visible.