Shift and Capacity Planning in a Call Centre

Call centre capacity planning: volume forecasting, service level targets, occupancy, shrinkage and shift design.

“How many agents do we need?” has no single answer but one method: first decide what you are promising, then calculate the headcount that keeps the promise. It does not work in the other order.

1. Set a service level target

Capacity planning starts with a target. For example: “80% of calls answered within 20 seconds.” Without that sentence no calculation is possible, because the required headcount depends entirely on it. Raising the target (say to 90% in 10 seconds) increases headcount disproportionately.

2. Forecast volume by time slot

A daily total is not enough for planning. You need it broken down by:

  • Day of week (Monday is usually the busiest).
  • Hour of day (most businesses peak twice, around 10:00–12:00 and 14:00–16:00).
  • Seasonal effects (campaign periods, billing days, the day before a holiday).
  • Channel mix (phone and messaging have different load curves).

Eight to twelve weeks of history is the minimum for a forecast you can trust.

3. Calculate the workload

A simple multiplication: calls per hour × average handle time = total talk minutes needed in that hour. For example 120 calls at 4 minutes AHT gives 480 minutes of workload, or theoretically 8 agent-hours. But that raw figure understates the real requirement substantially.

4. Add shrinkage

An agent cannot be on calls for the whole shift. What has to be accounted for:

  • Breaks and meals
  • Training and meetings
  • After-call work (logging, note-taking)
  • Leave and absence
  • System or technical downtime

In practice total shrinkage runs between 25% and 35%. Eight raw agent-hours becomes roughly 11–12 people at 30% shrinkage.

5. Keep occupancy reasonable

Occupancy is the share of an agent’s logged-in time spent on calls. Consistently above 90% looks efficient in the short term; in the medium term it produces burnout, more errors and staff turnover. The sustainable band is usually 75–85%.

Designing shifts

Staggered starts

Instead of everyone starting at 09:00, running 08:30/09:00/09:30 tiers smooths capacity across the peaks.

Part-time cover

Short shifts aimed at peak hours can be more economical than full-time headcount.

Cross-training

An agent who knows more than one topic can be redirected when load shifts.

Break planning

If every break lands between 12:00 and 13:00, service level collapses in that hour. Distribute breaks against volume.

The weekly review

  1. Compare last week’s forecast with what actually happened; if the deviation keeps pointing the same way, fix the model.
  2. Find the time slot where service level breaks down most.
  3. Look at shrinkage in that slot — the problem is usually the break and meeting plan rather than headcount.
  4. Make one change for the following week and measure its effect.

Frequently asked questions

Are these calculations meaningful for a small team?

The formulas are more accurate at volume, but the logic holds at any size: set a target, add shrinkage, keep occupancy sustainable.

How do I plan for messaging channels?

In messaging one agent can run several conversations at once. Calculate capacity with a concurrency factor rather than the call formula.

Is abandonment part of the target?

It is usually tracked separately but should be read together: if service level looks fine while abandonment is high, people are hanging up before being counted.

Next step: to see how hourly call reports come out, request a demo.