What is call center shrinkage?

Call center shrinkage is the percentage of paid time that team members are not available to handle customer contacts. It covers everything from scheduled breaks, training, coaching and meetings through to unplanned absence, late starts, and system downtime.

Shrinkage is the number that turns a volume forecast into a staffing requirement. If a forecast says a contact center needs 100 people handling conversations at 10am, and shrinkage runs at 30%, then roughly 143 people need to be on the payroll to put 100 on the floor. Getting shrinkage wrong is the most common reason a well-forecast center still misses its service targets.

This page covers what counts as shrinkage, the planned and unplanned categories, how to calculate it, what a typical rate looks like, how it feeds the staffing calculation, and what can and cannot be reduced.

Call center shrinkage in one sentence

Shrinkage is the gap between the hours you pay for and the hours available for customer contact.

What counts as call center shrinkage

Shrinkage is usually split into two categories, and the split matters because they respond to completely different management.

• Planned shrinkage is known in advance and appears in the schedule: breaks, lunch, team meetings, training, coaching and one-to-ones, approved holiday, and any scheduled project work.

• Unplanned shrinkage is not known in advance: sickness, late arrivals, early departures, unapproved absence, system outages, and time lost to technical faults.

Some organizations add a third category for the small, hard-to-schedule losses that sit between the two, such as time spent walking to a meeting room or waiting for a supervisor. These are real and add up, and they are frequently the difference between a shrinkage model that works and one that consistently under-staffs.

The category split is worth maintaining. Planned shrinkage is a budgeting question. Unplanned shrinkage is an operational one. A center at 32% shrinkage that is almost all planned is in a very different position from one at 32% that is running heavy on unplanned absence, even though the headline figure is identical.

What is not shrinkage

After-contact work is not shrinkage. Someone writing up notes after a conversation is doing customer work and their time counts as handle time. This gets miscategorized often enough to be worth stating.

Idle time waiting for the next contact is also not shrinkage. That is available time, and it belongs in the occupancy calculation instead.

How to calculate call center shrinkage

Shrinkage = total unavailable hours ÷ total paid hours × 100

Both figures come from the same period. Unavailable hours are planned plus unplanned shrinkage combined.

A worked example. A team of 50 people is paid for 2,000 hours in a week. Across that week, 300 hours go to breaks and lunch, 120 to training and meetings, 90 to holiday, and 130 to sickness and lateness. Total unavailable is 640 hours.

Shrinkage = 640 ÷ 2,000 × 100 = 32%

Using shrinkage to staff

The reason to calculate shrinkage at all is to get from a required-on-floor number to a required-on-payroll number:

Staff required = base requirement ÷ (1 − shrinkage)

Using the 32% above, a base requirement of 100 people on the floor becomes 100 ÷ 0.68, or roughly 147 people scheduled.

This is where shrinkage errors get expensive. Understating shrinkage by five points on a 100-person requirement leaves you about eleven people short every day, which shows up as missed service level rather than as a planning error, and usually gets blamed on the wrong thing.

The details that decide the number

The denominator. Paid hours and scheduled hours are not the same thing, and using them interchangeably produces figures that cannot be compared week to week. Holiday in particular sits inside paid hours and outside scheduled hours, which moves the result by several points on its own.

The period. Weekly shrinkage swings hard with holiday and training cycles. Monthly or quarterly figures are more stable for planning, but interval-level shrinkage is what you need for intraday staffing, since breaks cluster and training is rarely spread evenly across the day.

Whether holiday is in or out. Some centers exclude annual leave from shrinkage and handle it separately in the annual capacity plan. That is a legitimate approach and produces a much lower headline figure. It also makes the number incomparable with anyone else's, so document the choice.

What is a typical call center shrinkage rate?

Most contact center professionals put shrinkage between 30% and 35%. Dimension Data's Global Benchmarking Report gives an average of 35%. Call Centre Helper's own worked example, built component by component from annual leave through to system problems, totals 32.5%, while the average figure planners actually enter into its Erlang calculator is 26.6%.

That spread between what gets modelled and what the component arithmetic produces is itself informative. It usually means uncoded activities are missing from somebody's plan.

That spread is wide enough that the benchmark is close to useless as a target. What matters far more is whether your own figure is accurate and stable, because the staffing calculation depends on it. A center that knows it runs at 38% and plans for 38% is in better shape than one that assumes the industry-standard 30% and is short every day.

Structural factors that legitimately push the number up: regulated industries with heavy compliance training, teams handling complex products that need continuous coaching, seasonal operations with large onboarding cohorts, and any center where the tooling takes a long time to load.

Shrinkage vs. occupancy vs. utilization vs. adherence

These four measure different slices of the same paid hour.

The clean way to hold them apart: shrinkage tells you how many people to hire, occupancy tells you whether you hired the right number, adherence tells you whether the ones you have are where the plan says they should be.

Why shrinkage matters

It is the bridge between forecast and headcount. No other metric in workforce management has as direct an effect on whether a center is staffed correctly. Every other planning number depends on it being right.

It makes hidden costs visible. Training, coaching, and meetings are all necessary and all consume paid hours that cannot handle contacts. Tracking them as shrinkage means those decisions get made with the cost in view rather than absorbed silently into a service level miss.

It separates a people problem from a planning problem. Splitting planned from unplanned shrinkage tells you immediately whether a coverage gap comes from over-scheduling activities or from absence. Those need entirely different responses.

Where shrinkage gets misused

Treating it as waste to eliminate. A good chunk of shrinkage is training, coaching, and breaks. Cutting those to improve the number produces an undertrained, exhausted team and a worse operation. The objective is an accurate shrinkage figure, not a low one.

Applying an industry benchmark instead of measuring. The 30% convention is an average across wildly different operations. Planning against someone else's number rather than your own is the most common way centers end up structurally short.

Ignoring the interval view. Shrinkage is not evenly distributed. Breaks cluster around the middle of the day, training tends to sit in the afternoon, and absence peaks on Mondays. A flat daily rate applied evenly across intervals will over-staff some and under-staff others.

Blending planned and unplanned. The headline figure hides which of the two is moving. Track them separately or the number tells you nothing actionable.

How to manage call center shrinkage

The realistic target is accuracy first, then reduction where reduction is genuinely available.

Measure what you actually have. Most centers discover their real shrinkage is higher than the figure in the plan, usually because uncoded activities were never captured. Fixing the measurement is worth more than any reduction effort.

Schedule the planned portion deliberately. Training and meetings placed in the quietest intervals cost far less coverage than the same hours placed at peak. This is the single highest-return shrinkage action and it does not reduce the number at all.

Address the causes of unplanned absence rather than the absence. Sustained high occupancy, poor scheduling, and inadequate tooling all show up later as sickness and attrition. Unplanned shrinkage is frequently a symptom of something upstream.

Cut the mechanical losses. Login sequences, slow-loading applications, and manual handoffs between systems consume paid minutes every single day per person. A workspace that opens once and holds everything in one place removes a recurring loss that never appears on anyone's improvement plan.

Rebuild the model when the operation changes. Shrinkage assumptions age. A center that has added channels, changed its training programme, or shifted its staffing mix in the last year is almost certainly planning against a stale figure.

Shrinkage when AI handles part of the volume

Two effects, one of which is easy to miss.

The visible one: with fewer contacts reaching people, the base requirement falls, so the same shrinkage percentage translates into a smaller absolute number of scheduled hours. The arithmetic is unchanged.

The one that gets missed: shrinkage percentage itself tends to rise. Training load goes up rather than down, because the work reaching people is more complex and the tooling changes more often. Coaching becomes more valuable per hour and therefore more frequent. Meanwhile the base of paid hours the shrinkage is measured against has shrunk, so the same absolute training hours make a bigger percentage.

A center that keeps planning against its pre-AI shrinkage assumption will be short. The number needs recalculating whenever the contact mix changes materially, not once a year.