What Is Call Center Shrinkage? Formula, Examples and Benchmarks

Shrinkage is the gap between the people you pay and the people actually on the phones. A forecast or an Erlang C model tells you how many agents must be logged in and ready in each interval. Shrinkage tells you how many more you must schedule so that, after breaks, training, holidays and sick days, that many are really there.

Get it wrong and every other plan fails quietly: service level misses in the busiest intervals, occupancy climbs, agents burn out, and absence rises, which raises shrinkage further.

How do you calculate call center shrinkage?

Shrinkage = (scheduled hours − available hours) ÷ scheduled hours × 100.

Scheduled hours are the paid hours in the rota. Available hours are the hours agents were logged in and ready for contacts, including talk, hold, wrap-up and time waiting for the next contact. Everything else is shrinkage. You can calculate it per agent, per team, per day or per interval; for staffing, the per-interval figure is the one that matters.

Worked example: one agent's 40-hour week

This is an illustrative example, averaged over a year so that annual leave and public holidays are spread across every week.

Scheduled hours are 40, available hours are 40 − 12 = 28, so shrinkage is 12 ÷ 40 = 30%. Unpaid lunch breaks are left out on both sides: they are not in the paid hours, so they are not shrinkage.

Splitting the total this way is useful because the two halves are managed differently. Planned shrinkage is a scheduling decision. Unplanned shrinkage is a symptom, and the causes behind it can usually be found and reduced.

What is a typical call center shrinkage?

Call Centre Helper puts the industry average at 30% to 35%, and cites Dimension Data's Global Benchmarking Report for an average of 35% (checked October 2026). Treat that as a sanity check, not a target. A centre with generous leave, heavy onboarding or long training will run higher; a mature team with stable staff may run lower. If you do not have a figure yet, plan with 30% and replace it with your own measurement after a few weeks of real schedule and activity data.

Shrinkage vs occupancy vs utilisation

These three are often mixed up because they all describe how agent time is spent. They measure different slices of it.

The three are linked: utilisation = (1 − shrinkage) × occupancy. With 30% shrinkage and 85% occupancy, only about 60% of paid hours go on handling contacts. That is why cost models built on talk time alone undercount labour; the call center cost calculator works from paid hours for the same reason. Some teams define utilisation differently, so check the definition before comparing numbers between sites or vendors.

How does shrinkage feed into staffing?

Staffing runs in two steps. First, an Erlang C calculator turns calls, handle time and a service level target into the number of agents who must be on the phones in each interval. Second, shrinkage turns that into the number of people to schedule:

People to schedule = agents needed ÷ (1 − shrinkage), rounded up.

If Erlang C says an interval needs 18 agents and shrinkage is 30%, schedule 18 ÷ (1 − 0.30) = 18 ÷ 0.70 = 25.7, so 26 people. Check it: 30% of 26 is 7.8, leaving 18.2 available.

The common mistake is adding the percentage instead: 18 × 1.30 = 23.4, so 24 people. After 30% shrinkage only 16.8 are available, and the interval runs short by more than one agent all day. The Erlang C calculator has a shrinkage input that applies the division for you, and the guide how many call center agents do I need walks through the whole method.

Shrinkage also varies through the day and week. Breaks and meetings cluster mid-shift, training lands on quieter days, and absence often peaks on particular weekdays. If you can, plan with a shrinkage figure per interval rather than one weekly average.

How do you reduce unplanned shrinkage?

• Measure it by cause. Record absence, lateness, break overruns and downtime separately. A single shrinkage number hides which one is growing.
• Track schedule adherence. Compare what each agent was scheduled to do with what they actually did, interval by interval. Most lateness and overrunning breaks show up here first.
• Look for patterns in absence. Absence that clusters on certain days, shifts or teams usually has a cause: a rota that is hard to keep, a difficult campaign or a team-lead problem.
• Make time off easy to request early. Leave booked in advance becomes planned shrinkage you can schedule around; the same day taken at short notice becomes unplanned.
• Keep occupancy below its cap. Agents run near 100% busy get tired and take more sick days. Holding occupancy around 85% to 90% protects attendance.
• Fix system downtime at the source. Slow logins, softphone drop-outs and CRM outages are shrinkage the agent cannot control. Log every minute of it so it gets fixed.
• Do not cut planned shrinkage to hit a number. Removing training and coaching lowers shrinkage this month and raises errors and attrition later.

Shrinkage belongs alongside the other operational metrics in your reporting; the guide to call center KPIs and formulas shows how it sits with service level, handle time and occupancy.

How does DialerBee help with shrinkage?

DialerBee includes workforce management with shifts, time-off, schedule adherence and forecasting, inside the same platform as your queues. That puts the scheduled side of the shrinkage formula and the actual side in one place.

On the floor, supervisor tools give a live agent grid with each agent's state and how long they have been in it, and agent activity timers for time in ready, in wrap-up, on break and on a call. Coaching rules raise a supervisor's attention against agent activity, so an overrunning break or a long wrap-up surfaces on its own. Agent activity and utilisation reports show how a shift was actually spent.

For routine calls, AI voice agents and an AI receptionist can take routine calls on a number routed to them, on the same platform as your queues, and hand a call they cannot finish to a person. If you use one, model the calls it takes alongside your human staffing plan. Plans and prices are on the pricing page.

Frequently Asked Questions

What is shrinkage in a call center?

Shrinkage is the share of paid, scheduled agent time in which agents are not available to handle contacts. It covers planned time off the phones, such as breaks, training, meetings, coaching and annual leave, and unplanned time, such as sickness, lateness and system downtime.

What is the call center shrinkage formula?

Shrinkage = (scheduled hours − available hours) ÷ scheduled hours × 100. Available hours are the hours agents were actually logged in and ready to handle contacts. If an agent is paid for 40 hours and is available for 28, shrinkage is 12 ÷ 40 = 30%.

What is a typical call center shrinkage percentage?

Call Centre Helper puts the industry average at 30% to 35% and cites Dimension Data's Global Benchmarking Report for an average of 35%. Your own figure depends on leave policy, training load and absence, so measure it over several weeks rather than copying a benchmark.

How do I use shrinkage to work out how many agents to schedule?

Divide the agents you need on the phones by (1 − shrinkage) and round up. If Erlang C says 18 agents are needed in an interval and shrinkage is 30%, schedule 18 ÷ 0.70 = 25.7, so 26 people. Do not add 30% to 18: that gives 24 and leaves the interval two people short.

What is the difference between shrinkage and occupancy?

Shrinkage measures how much paid time is lost before agents are even available. Occupancy measures how much of the available time is spent handling contacts rather than waiting. Shrinkage is about time off the phones; occupancy is about how busy agents are while on them.

What is the difference between planned and unplanned shrinkage?

Planned shrinkage is time you schedule in advance: breaks, training, team meetings, coaching, annual leave and public holidays. Unplanned shrinkage is time you cannot schedule: sickness, short-notice absence, lateness, overrunning breaks and system downtime. Planned shrinkage is managed in the schedule; unplanned shrinkage is managed by finding and fixing its causes.

Is lower shrinkage always better?

No. Cutting training, coaching and breaks lowers shrinkage on paper but usually raises errors, absence and attrition later, which pushes unplanned shrinkage back up. The aim is to plan the shrinkage you need and reduce the shrinkage you do not.

Does DialerBee track shrinkage?

DialerBee includes workforce management with shifts, time-off, schedule adherence and forecasting, plus agent activity timers that show how long each agent spends in ready, wrap-up, on break and on a call. Those are the inputs you need to measure shrinkage and adherence from real data.

Related terms: agent occupancy, service level, average handle time, after-call work and wallboard.