Call Center Shrinkage Calculator
Work out what share of paid agent time is not available for contacts, split it into internal and external, planned and unplanned, and turn it into the number of agents you need to schedule.
Enter your team size and paid hours, then the hours lost to each shrinkage category. The calculator opens with a 50-agent example month. Staffing and cost fields are optional.
Hours lost to shrinkage
Internal (on site, off queue)
External (paid, not present)
Team total for the month. Leave a category blank if it does not apply.
Optional: staffing and cost
Total shrinkage for the month
31.0%
2,604 hours lost ÷ 8,400 paid hours × 100
Internal
20.5%
On site, off queue
External
10.5%
Paid, not present
Planned
26.0%
Known when scheduling
Unplanned
5.0%
Happens on the day
Effective agents
34.5
of 50 paid agents
Agents to schedule
56
38 required ÷ (1 − 31.0%)
Breakdown by category (% of paid hours)
- Paid breaks and lunch12.5%
- Coaching and one-to-ones2.5%
- Team meetings and huddles2.0%
- Training3.0%
- System downtime0.5%
- Vacation and holidays6.0%
- Sick and unplanned absence3.5%
- Late arrivals and early exits1.0%
Blue bars are internal categories; grey bars are external.
Tracking shrinkage over time? Download the free call center shrinkage report template to record it by category, week by week.
Stop estimating lost hours from memory. HiveDesk compares scheduled and actual hours for every shift and reports shrinkage hours and percentage by agent and date.
Track it automatically →How to Calculate Call Center Shrinkage
Shrinkage is the share of paid agent time that is not available for handling customer contacts. A center that pays for 2,000 agent hours a week never gets 2,000 hours of phone time: breaks, coaching, training, vacations, sick days, and outages all take a slice. Shrinkage measures the size of that slice, and it is the number that converts a forecast of agents needed into a schedule of agents to roster.
The calculation is a single ratio. The work is in collecting the inputs consistently, which is why the calculator asks for each category separately rather than one total.
The Shrinkage Formula
Paid hours = Agents × Paid hours per agent in the period
Shrinkage = Hours not available for contacts ÷ Paid hours × 100
Agents to schedule = Agents required on queue ÷ (1 − Shrinkage), rounded up
Worked example: a 50-agent team, one month
The calculator opens with this scenario, the same one used in our shrinkage glossary entry. Fifty agents each work 8 paid hours a day for 21 days, so the team is paid for 50 × 168 = 8,400 hours. During the month the hours lost to each category were:
| Category | Type | Hours | % of paid |
|---|---|---|---|
| Paid breaks and lunch | internal, planned | 1,050 | 12.5% |
| Coaching and one-to-ones | internal, planned | 210 | 2.5% |
| Team meetings and huddles | internal, planned | 168 | 2.0% |
| Training | internal, planned | 252 | 3.0% |
| System downtime | internal, unplanned | 42 | 0.5% |
| Vacation and holidays | external, planned | 504 | 6.0% |
| Sick and unplanned absence | external, unplanned | 294 | 3.5% |
| Late arrivals and early exits | external, unplanned | 84 | 1.0% |
| Total unavailable | 2,604 | 31.0% | |
- Total shrinkage = 2,604 ÷ 8,400 × 100 = 31.0%
- Internal shrinkage = 1,722 ÷ 8,400 × 100 = 20.5%; external = 882 ÷ 8,400 × 100 = 10.5%
- Planned shrinkage = 26.0%; unplanned = 5.0%
- Effective agents = 50 × (1 − 31.0%) = 34.5 agent-equivalents available for contacts
In other words, the center is paying for 50 agents and getting the contact-handling capacity of about 34. That is not a sign of a badly run team. Most of the gap is breaks, leave, and training that any responsible employer provides. The point of the calculation is to know the number precisely, so the schedule is built around it instead of discovering it when the queue backs up.
Internal vs External, Planned vs Unplanned
A total shrinkage rate tells you how big the problem is. Splitting it tells you who can fix it. Planners use two independent splits, and the calculator reports both.
Internal shrinkage
Paid time when the agent is at work but not available for contacts: paid breaks, coaching, team meetings, training, and system downtime. It is largely controllable through scheduling. Coaching moved from a peak Monday morning to a quiet Wednesday afternoon still counts as shrinkage, but it stops costing service level. Internal shrinkage shows up alongside schedule adherence, because an agent taking a 25-minute break against a 15-minute plan is both non-adherent and adding shrinkage.
External shrinkage
Paid time when the agent is not at work: vacation, public holidays, sick leave, and the minutes lost to late arrivals and early departures. It is managed through leave planning and attendance policy rather than the intraday schedule. Our guide to absenteeism covers how to measure the unplanned part.
Planned vs unplanned
The second split asks whether you knew about the lost time when the schedule was built. Approved vacation, scheduled training, and paid breaks are planned, so they can be placed in low-volume intervals. Sick calls, no-shows, late logins, and outages are unplanned, so they must be covered by a buffer of extra scheduled agents. A center can have high planned shrinkage and still hit service level; high unplanned shrinkage almost always hurts, because nobody can schedule around it.
The two splits overlap. Vacation is external and planned; system downtime is internal and unplanned. The calculator tags each category both ways, so you can see, for example, that most of your unplanned shrinkage comes from sick leave rather than lateness, and decide whether the fix is an attendance policy or a coverage plan.
Turning Shrinkage into a Schedule
Queueing models such as the Erlang C calculator tell you how many agents must be logged in and available during an interval to meet a service level target. They assume every scheduled agent is available. Shrinkage closes the gap between that assumption and reality.
The correct adjustment is to divide by the share of time that remains, not to add the shrinkage percentage on top. In the worked example, Erlang C says 38 agents are needed on the queue and shrinkage is 31.0%:
- Correct: 38 ÷ (1 − 0.31) = 55.1, so schedule 56 agents.
- Common mistake: 38 × (1 + 0.31) = 49.8, so 50 agents, which is 6 short.
The error grows as shrinkage rises. The table below holds the on-queue requirement at 38 agents and shows the gap at different shrinkage levels.
| Shrinkage | Schedule (÷ method) | Schedule (× method) | Shortfall |
|---|---|---|---|
| 20% | 48 | 46 | 2 |
| 25% | 51 | 48 | 3 |
| 30% | 55 | 50 | 5 |
| 35% | 59 | 52 | 7 |
| 40% | 64 | 54 | 10 |
Agents required on queue: 38. Both methods rounded up to whole agents.
Apply shrinkage interval by interval where you can, not as one daily average. Breaks and coaching cluster at particular times of day, so a 31% daily average might be 20% at 9 a.m. and 40% after lunch. For hiring plans, convert the scheduled requirement into headcount with the FTE calculator, since part-time agents and overtime change how many people that schedule needs.
What Counts as Shrinkage (and What Does Not)
Most disagreements about shrinkage rates are really disagreements about definitions. Before comparing your number with a benchmark or another site, check these boundaries.
| Time | Shrinkage? | Why |
|---|---|---|
| Paid breaks | Yes | Paid time, not available for contacts. |
| Unpaid lunch | No | Not in paid hours, so outside the ratio entirely. |
| Coaching, meetings, training | Yes | Paid and off queue, even when valuable. |
| After-call work | No | Part of handle time; the agent is still working a contact. |
| Idle time while available | No | Measured by occupancy, not shrinkage. |
| System downtime | Yes | Agents are paid but cannot take contacts. |
| Vacation and holidays | Yes | Paid leave reduces available hours. |
| Unpaid leave | No | Not paid, so not in the denominator. |
| Late logins and early exits | Yes | Paid shift time lost to attendance. |
| Off-queue project work | Usually | Count it if the agent is paid and unavailable for contacts. |
The two most common mistakes are counting idle time and counting after-call work. Both are already built into the Erlang model through occupancy and average handle time, so adding them to shrinkage counts the same minutes twice and leads to overstaffing.
Paid hours should include overtime actually worked, because those are hours you bought. If some agents work extra shifts to cover absences, their overtime adds to the denominator while the absence adds to the numerator, and the rate reflects what really happened. Whatever rules you choose, write them down and apply them every period; a definition that changes month to month makes the trend meaningless.
Shrinkage Benchmarks and Cost
Many workforce planners use a starting assumption of around 30% shrinkage when they have no history of their own. Treat that as a placeholder, not a target. Published benchmarks vary widely, and they rarely say whether they include paid leave, how they treat lunch, or whether they cover inbound voice only. A center that gives generous paid vacation and weekly coaching can run well above 30% and still be well managed; a center at 22% may simply not be training anyone.
The better benchmark is your own history, split by category. Watch the unplanned share in particular. If planned shrinkage is steady and unplanned shrinkage is creeping up month after month, the cause is usually attendance, morale, or scheduling, and it is worth investigating before it turns into turnover.
Cost example
Suppose the loaded cost of an agent (wage plus payroll taxes and benefits) is $22 an hour, a hypothetical figure for illustration. In the worked example, 2,604 paid hours were not available for contacts, which is $57,288 for the month and $687,456 a year if the month is typical. Use the employee cost calculator to work out your own loaded hourly cost.
That figure is not waste. Most of it buys breaks, leave, and training you would not want to cut. Its use is in pricing the parts you can influence: if unplanned shrinkage is 5.0% of paid hours, each percentage point of it in this example is worth about $1,848 a month, which puts a number on the value of a better attendance policy or a more reliable phone system.
How to Measure and Reduce Shrinkage
The calculator works from numbers you already have. The harder problem is getting those numbers without a supervisor reconstructing them from spreadsheets each month. These are the practices that make shrinkage measurable, and the HiveDesk features that support each one.
Track off-queue time by category
Set up separate categories for queues, training, coaching, meetings, and admin, and have agents track time against each one. In HiveDesk these are projects, and time reports show the breakdown by agent and category. Break types with short codes separate paid breaks from other off-queue time.
Compare attendance against the schedule
External unplanned shrinkage lives in the gap between the shift and the clock-in. HiveDesk compares check-in and check-out times with each scheduled shift, applies a late threshold you set, and shows who is present, absent, on leave, or not logged in on the Scheduled Activity view. The Adherence report summarizes check-in and check-out adherence by site. See attendance management.
Plan leave and training into the schedule
Planned shrinkage only hurts when it lands in the wrong interval. Record approved time off in leave management so it is visible when shifts are built, and place training and coaching in your lowest-volume periods. Our guide to shift scheduling and shrinkage management covers the scheduling side in more depth.
Review the Shrinkage report weekly
HiveDesk's Shrinkage report compares each agent's scheduled hours with actual hours worked and shows shrinkage hours and shrinkage percentage by agent and date. It filters by site, team, or agent and exports to PDF or Excel. Because it works from scheduled versus worked time, it captures attendance-driven shrinkage directly; combine it with time tracked against training and meeting categories for the full picture. A weekly review catches patterns, such as the same team losing time every Monday, while they are still easy to fix. See shrinkage management.
HiveDesk is a time tracking, attendance, and scheduling tool. It does not forecast contact volume, run Erlang calculations inside the product, or connect to your phone system for real-time queue states. It gives you the tracked hours, attendance, and schedule data that a shrinkage calculation depends on, for $5 per user per month with every feature included. For background on the full topic, read our article on understanding and fixing contact center shrinkage.
Frequently Asked Questions
Add up every paid hour in the period when agents were not available to handle contacts (paid breaks, coaching, meetings, training, system downtime, vacation, sick leave, and lateness), divide by total paid hours, and multiply by 100. For example, 2,604 unavailable hours out of 8,400 paid hours is 2,604 ÷ 8,400 × 100 = 31.0% shrinkage.
Internal shrinkage is paid time when the agent is at work but off the queue: breaks, coaching, meetings, training, and system downtime. External shrinkage is paid time when the agent is not at work at all: vacation, public holidays, sick days, and lateness. Internal shrinkage is managed mainly through scheduling; external shrinkage through leave planning and attendance management.
Planned shrinkage is known when the schedule is built, such as approved vacation, scheduled training, and paid breaks, so it can be placed in low-volume intervals. Unplanned shrinkage happens on the day, such as sick calls, no-shows, late logins, and outages, so it has to be covered by a buffer of extra scheduled agents. The planned and unplanned split is the more useful one for intraday management.
There is no single good number, because shrinkage depends on how much training, coaching, and paid leave a center offers, and benchmarks use different definitions. Many planners start with an assumption of around 30% when they have no history, then replace it with their own measured figure. The most useful target is a stable, well-understood rate with a falling unplanned share, not the lowest possible total.
Divide the number of agents you need on the queue by one minus the shrinkage rate, then round up. If an Erlang C calculation says you need 38 agents available and shrinkage is 31%, you schedule 38 ÷ 0.69 = 55.1, so 56 agents. Multiplying by 1.31 instead gives 50 and leaves the interval understaffed.
No. After-call work is part of average handle time, because the agent is still working a contact. Idle time between contacts while logged in and available is measured by occupancy, not shrinkage. Shrinkage only covers time when the agent is unavailable to take a contact at all. Counting either of them as shrinkage double-counts time that the Erlang model already accounts for.
No. Shrinkage is measured against paid hours, so unpaid time is outside the calculation altogether. If agents work an 8.5-hour shift with a 30-minute unpaid lunch, paid hours are 8 per shift and only the paid 15-minute breaks count as shrinkage. Mixing paid and unpaid time is one of the most common reasons two centers report very different rates for similar operations.
Measure it continuously and review it at several levels. Look at it by interval and day of week to place coaching and training, by month to catch seasonal patterns such as summer vacations and winter sickness, and as a rolling 12-month average for hiring plans. A single annual figure hides the weeks when shrinkage is high enough to break service level.
Measure Shrinkage from Real Hours, Not Estimates
HiveDesk tracks time by activity, compares attendance against every shift, and reports shrinkage by agent and date. $5/user/month, 14-day free trial, no credit card.
Ready to Get Started?
Join teams worldwide who trust HiveDesk for workforce management, time tracking, and employee monitoring. $5/user/month, all features included.