How to Reduce Call Center Shrinkage Without Policing Your Team

Every vendor answer to how to reduce call center shrinkage ends with buying their software, and every honest answer ends on your own floor.

I sat through a capacity review in 2021 where our WFM manager put up a slide showing shrinkage at 34 percent against a planned 30. Four points. On a 240 seat account that is roughly ten heads we were paying for and not getting. The room spent forty minutes discussing whether to procure a new adherence tool.

Not one minute was spent on the fact that my Bengaluru team's average break overrun was six minutes a day.

Shrinkage in a contact center is the percentage of paid agent time that is not available to take contacts. It covers breaks, meetings, training, coaching, system downtime, after call work in some models, and every category of absence. It is the single number that decides how many people your account has to hire to answer the same volume, and it is the number most team leaders have never been shown for their own team.

What Is Shrinkage in a BPO, and Why Your WFM Team Keeps Raising It

Shrinkage is the gap between paid hours and productive hours. If you pay for 100 hours and get 70 hours of agents actually available to handle contacts, your shrinkage is 30 percent, and WFM has to staff for it by hiring more people than the volume alone would suggest.

The question of what is shrinkage in a BPO usually gets answered with a definition and stops there. The useful part is the split, because the two halves behave completely differently.

Planned shrinkage is what you scheduled: training, team meetings, coaching sessions, annual leave approved in advance. WFM can plan around every minute of it and it costs almost nothing in service level.

Unplanned shrinkage is what happened to you: sick calls, no shows, late logins, system outages, break overruns, unlogged aux time. This is the expensive half, because capacity was committed against it and nobody can recover the interval.

Most accounts run planned shrinkage between 12 and 18 percent and unplanned between 8 and 16 percent. When a client complains about shrinkage, they are almost always looking at the unplanned half, and almost all of the unplanned half is controlled at team leader level rather than in the WFM room.

That is the part that gets lost. WFM reports shrinkage. Team leaders create it.

Anyone asking what is shrinkage in a BPO for the first time should stop at the planned versus unplanned split and go no further until it is clear which half their own number sits in.

The Call Center Shrinkage Formula and the Number It Hides

The standard call center shrinkage formula is straightforward, and its simplicity is exactly what lets a problem hide inside it.

Shrinkage % = (Total paid hours minus Productive hours) ÷ Total paid hours × 100

Run through it with real numbers. A team of 20 agents on nine hour shifts across a five day week is 900 paid hours. If the login and aux data shows 612 hours of actual contact handling availability, shrinkage is 32 percent.

The trouble with a single blended number is that 32 percent tells you nothing about what to fix. A better version splits it by driver.

Add the four team leader rows at their midpoints and you land between 7 and 9 percent. On a 240 seat account, that is the difference between hiring 20 extra people and hiring none.

I run this split for any lead who tells me their shrinkage problem belongs to WFM. It takes one export and about thirty minutes, and it usually ends the argument.

The call center shrinkage formula gives you one figure. This table gives you a work list, and only one of the two is any use on a Monday morning.

How to Reduce Call Center Shrinkage in the Four Places It Actually Leaks

Work the four recoverable rows in order of size, not in order of how annoying they are. Attack all four at once and you get a team that feels surveilled and a number that moves for six weeks before returning to where it was.

Pick one per month. I have never seen the all at once approach hold.

Everything that follows is how to reduce call center shrinkage using changes the team either does not notice or actively welcomes, which is the only version that survives a quarter.

Break Creep: The Six Minutes Nobody Counts

Break creep is the gap between a 15 minute break and an 18 minute one, multiplied by every agent, every day. It is the largest of the four leaks and the least resented when handled properly, because almost nobody is doing it deliberately.

On my floor the causes were consistent and physical:

1. Queue at the canteen or the lift. Two minutes each way that no policy acknowledges.
2. Staggering that puts nine people on break at once. Everyone waits.
3. The walk itself. A 90 second walk each way inside a 15 minute break is a 12 minute break.
4. No clock the agent can see. Most floors have one clock and most agents cannot see it from the canteen.

The fixes are environmental rather than disciplinary. We restaggered breaks into three waves rather than two, which removed the canteen queue. We moved the break start to the moment the agent went aux rather than when they left the desk. Break overrun dropped from just over five minutes per agent per day to about ninety seconds in seven weeks, and nobody was warned or counselled.

That is roughly 1.4 percent of shrinkage recovered on a 20 person team, with zero conversations about discipline. It is the cleanest example I have of how to reduce call center shrinkage by changing the building rather than the people in it.

After Call Work Time Reduction Without Rushing the Customer

After call work, or ACW, is the time an agent spends wrapping up a contact before taking the next one. It is legitimate work and it is also the easiest place on the floor to hide. Real after call work time reduction comes from removing the reason for the padding, not from cutting the allowance.

Ask why the ACW is long before you touch it. On my accounts the answer was usually one of three things.

• The agent is taking a breath. After a difficult call this is healthy and the honest fix is a proper recovery aux code, not a shorter wrap.
• The system is slow or the notes field is badly built. Six clicks to close a case is a process problem that no coaching fixes.
• The disposition list has 40 entries and the agent cannot find the right one. I saw a floor where the average wrap fell by 22 seconds after we cut a disposition list from 41 codes to 12.

Genuine after call work time reduction usually comes from the second and third causes. When we rebuilt a case notes template on one process, average ACW went from 78 seconds to 51 without a single conversation with an agent about speed.

The failure mode in after call work time reduction is setting an ACW target and coaching against it. Agents hit the target by wrapping during the next call, which shifts the cost into quality and AHT where it is harder to see.

Unlogged Aux and the Codes People Use to Disappear

Every contact center has an aux code that has quietly become the place people go. Usually it is a training code, a system issue code, or a legacy project code nobody retired. It shows up as productive or neutral time in the report and it is not.

Find it by pulling aux usage by code by agent for four weeks and looking for the outlier. Not the outlier agent, the outlier code.

On one of my accounts, an aux code labelled "Process Clarification" was carrying 40 minutes a day across the team. It existed because a real process change three years earlier had genuinely needed it. It had become the smoke break code. Retiring it took one email and recovered about 1.2 percent of shrinkage.

You may be thinking this is exactly the surveillance I said to avoid. The difference matters: I looked at the code, not at the people, and I fixed the structure rather than counselling twelve agents. Nobody was named and nobody had to be.

How to Reduce Unplanned Absenteeism in a Call Center Without Attendance Policies

Unplanned absence is the hardest of the four and the only one where the cause sits outside the shift. Any approach to how to reduce unplanned absenteeism in a call center that starts with a stricter policy will produce a short dip and then a worse number, because the underlying reason has not moved.

Look at the distribution before you look at the policy. On every floor I have run, unplanned absence clustered in three places: Mondays, the day after payday, and the weeks following a roster change nobody explained.

What moved the number for me:

1. Call the agent on day one of absence, personally, not from HR. Not to check up. To ask if they are alright. This alone cut repeat absence noticeably.
2. Fix the Monday roster. We stopped scheduling the same people for Sunday and Monday back to back and Monday absence fell by roughly a third.
3. Publish the roster on a fixed day. Late rosters generate absence because people make plans anyway.
4. Ask the top three absentees what is happening, privately and without a warning letter in the room. Twice in my career it was a transport route change. Once it was a family illness nobody had told us about.
5. Stop rewarding perfect attendance with certificates. Reward the team, not the individual, or you create pressure to come in sick and infect the floor.

Nothing in that list is a policy change, which is the point. How to reduce unplanned absenteeism in a call center is almost always a scheduling and relationship question wearing an attendance label.

Most absence on a contact center floor is not defiance. It is a commute, a shift pattern, or a manager the agent is avoiding. When absence tracks to one team and not the account, the cause is the lead, and there is more on that pattern in my piece on what actually drives attrition in a call center.

Why Surveillance Always Backfires on Shrinkage

Tightening monitoring produces a fast improvement and a slow, expensive reversal. It works for about six weeks, which is long enough to look like a success in a monthly review and not long enough to show the cost.

I have made this mistake. In 2016 I put a daily adherence name and number list on the team board. Adherence improved four points in a month. Within a quarter I had lost two of my best agents, my coaching conversations had turned into negotiations about seconds, and agents had started staying logged in while doing nothing, which is worse than aux because it corrupts the data WFM plans from.

The pattern is consistent. Structural fixes hold and behavioural policing does not, because policing treats a floor design problem as a character problem. Every durable answer to how to reduce call center shrinkage sits in the bottom three rows of that table.

The International Customer Management Institute has published extensively on how adherence and occupancy targets affect agent behaviour, and the finding that matches my own experience is that agents optimise for whatever is measured most visibly, including in ways that damage the underlying number.

What to Do If Your Shrinkage Is Already Being Escalated

When a client or an ops head is already on your shrinkage number, the instinct is to announce a crackdown. Resist it for one week and do this instead.

• Day 1 to 2. Run the call center shrinkage formula for your own team, then pull the split by driver from the table above. Know which of the four rows is yours before anyone asks.
• Day 3. Take the split to your manager with the largest single recoverable row circled and a named fix. Owning the number changes the conversation completely.
• Day 4 to 5. Implement the one structural fix. Break staggering or an aux code retirement can both be done inside a week.
• Week 2 onward. Report the movement weekly, by driver, not as a blended percentage.

This is also, incidentally, how people get noticed for WFM roles. The analysis above is roughly what a scheduler does, and the route from the floor into that function is covered in the guide to the workforce management analyst career path.

Frequently Asked Questions About How to Reduce Call Center Shrinkage

These come up on every floor, usually the week after a capacity review goes badly.

How to reduce call center shrinkage when the number is already inside plan?
Stop working the total and start working the composition. An account at 29 percent against a 30 percent plan can still be carrying three points of break overrun offset by an underspend on training, which means you are quietly trading development time for availability. Split the number by driver quarterly even when the blended figure looks healthy.

What is a good shrinkage percentage for a call center?
Most well run accounts sit between 25 and 35 percent total, with planned shrinkage at 12 to 18 percent and unplanned at 8 to 16 percent. Anything under 25 percent usually means something is being excluded from the calculation rather than that the floor is unusually efficient. Compare against your own account's last four quarters rather than against an industry figure, because the categories included vary by client contract.

Does after call work count as shrinkage?
It depends on the contract, and this causes more arguments than any other line. Many models treat ACW as productive time because the agent is working on a contact, while others count it as shrinkage because the agent is unavailable for the queue. Check which model your account uses before you start any after call work time reduction work, because in one model the savings show up in shrinkage and in the other they show up in AHT.

What is the call center shrinkage formula in simple terms?
Take total paid hours, subtract the hours agents were actually available to handle contacts, divide by total paid hours, and multiply by 100. For a team of 20 on nine hour shifts over five days, that is 900 paid hours; if 612 were available, shrinkage is 32 percent. The formula is easy and the value only becomes useful once you split it by driver.

How to reduce call center shrinkage quickly without upsetting the team?
Retire misused aux codes and restagger breaks. Both are structural, both can be done in under a week, and neither requires a single conversation about individual behaviour. Between them they typically recover 2 to 3 percent, which is usually enough to change the tone of the next review while you work on the slower absence problem.

Is shrinkage a WFM responsibility or a team leader responsibility?
WFM measures and forecasts it. Team leaders create most of the recoverable portion. Planned shrinkage belongs to WFM and HR planning, while break overrun, aux misuse, ACW padding and unplanned absence, which together run 7 to 9 percent on most accounts, sit squarely with the floor.

How to reduce unplanned absenteeism in a call center when the commute is the problem?
Attack the roster rather than the attendance policy. Check whether your absent agents share a route or a shift pattern, then look at cab timings, back to back weekend scheduling, and shift start times against the transport window. On one of my accounts, shifting a single shift start by 30 minutes fixed an absence cluster that two rounds of warning letters had not touched.

Will an adherence tool fix my shrinkage?
It will measure it better, which is not the same thing. A tool tells you that break overrun is five minutes; it does not tell you that the canteen queue is three minutes of it. Buy the tool if you have no visibility at all, and do not expect the number to move until someone uses the output to change something physical on the floor.

What is shrinkage in a BPO compared to occupancy and utilisation?
Shrinkage is the share of paid time that is not available for contacts. Occupancy is how busy agents are during the time they are available. Utilisation blends the two into productive time over paid time. High occupancy with high shrinkage is the worst combination, because it means fewer agents are working harder, which is the exact pattern that produces burnout and then more absence.

How long before shrinkage improvements show up in the numbers?
Structural fixes like aux code retirement show in the following week. Break restaggering takes four to seven weeks because habits adjust slowly. Absence work takes a full quarter, since you are waiting on a behavioural pattern rather than a system change. Anything that improves in three days has usually moved into a different aux code rather than gone away.

Should I tell my team we are working on shrinkage?
Yes, and explain it in headcount rather than percentages. Saying "we lose the equivalent of two people a week to break overrun and late logins" is understood immediately, where "shrinkage is at 34 percent" is not. Teams cooperate with a problem they can picture and resist a metric that sounds like it exists to catch them.

Does remote or hybrid working increase shrinkage?
Unplanned absence usually falls and unlogged aux usually rises. The commute related absence largely disappears, while the informal time that used to be visible on a floor becomes invisible. The net effect on most accounts I have seen was close to neutral, with the composition changing rather than the total.

What is the one change with the highest return?
Retiring stale aux codes. It costs one email, takes no negotiation, recovers 1 to 3 percent on most floors, and it is the only item on the list where the agents affected generally do not notice or object. Start there, then move to break staggering.

The Number Is a Floor Problem With a Spreadsheet Name

Learning how to reduce call center shrinkage is mostly learning to stop treating it as a WFM report and start treating it as four specific, physical things happening on your floor every day. A canteen queue. A disposition list with 41 entries. An aux code that outlived its purpose. A Monday roster that nobody explained.

None of those get fixed by a procurement decision, and all of them get worse under surveillance. The call center shrinkage formula will never point at any of them, because a formula cannot see a lift queue.

That 2021 capacity review ended with an approval to evaluate a new adherence tool. We never bought it. We restaggered breaks, killed two aux codes, and rebuilt a notes template, and the number came back inside plan by the following quarter.

The four points we were missing were never hiding in the software. They were sitting six minutes at a time in a lift queue that nobody with budget authority had ever stood in.