A floor that looks fully staffed on the roster can still run short by mid-morning. The cause is rarely a forecasting error. It is usually time that was paid for but never available to take calls.
That gap is called shrinkage, and the BPO shrinkage formula is how it gets measured. Without it, a roster is built on hours that do not exist.
Call Centre Helper reports that most contact centre professionals see shrinkage land between 30 and 35 percent, and cites Dimension Data’s benchmarking report for an industry average of 35 percent.
This guide covers the BPO shrinkage formula, a worked example, the staffing conversion, the mistakes that distort the number, and how AI changes the calculation.
TL;DR
- The BPO shrinkage formula is Shrinkage % = (external shrinkage hours + internal shrinkage hours) / total paid hours x 100.
- In the worked example, 4,720 unproductive hours out of 16,000 paid hours gives 29.5 percent shrinkage, so about 70.5 of every 100 paid agents are available for customer contacts.
- Most contact centre professionals see shrinkage between 30 and 35 percent, and Dimension Data’s benchmarking report gives an average of 35 percent.
- Required roster = required productive agents / (1 – shrinkage), so 100 productive agents need about 142 rostered agents at 29.5 percent shrinkage.
- Adding the shrinkage percentage to the requirement instead of dividing leaves about 8 seats uncovered in that same example.
- AI voice agents remove breaks, leave, and coaching from the share of work they handle, but human rosters still need shrinkage planning for every remaining call.
What Is Shrinkage in a BPO?
Shrinkage in a BPO is the share of paid hours in which agents cannot handle customer contacts, for any planned or unplanned reason. Workforce planners use it as an allowance between the staff a forecast says are needed and the staff actually available at a given time.
Planners usually classify it in two ways. External shrinkage covers time lost outside the centre, such as leave, sickness, lateness, and paid breaks. Internal shrinkage covers time lost inside the centre, such as team meetings, coaching, training, and system downtime.
The second lens splits shrinkage into planned and unplanned. Planned and unplanned shrinkage need different controls, because a scheduled training session can be moved while an unexpected outage cannot.
The table below maps common components to both lenses.
| Component | Where It Happens | Planned or Unplanned |
|---|---|---|
| Annual leave and public holidays | External | Planned |
| Sickness and absenteeism | External | Unplanned |
| Lateness | External | Unplanned |
| Paid breaks | External | Planned |
| Team meetings and one-to-ones | Internal | Planned |
| Coaching and training | Internal | Planned |
| System downtime | Internal | Unplanned |
| Special projects and helping other teams | Internal | Varies |
Coaching and quality sessions count as internal shrinkage even though they lift performance. That is why BPO quality parameters and shrinkage targets should be planned together instead of traded against each other.
What Is the BPO Shrinkage Formula?
The BPO shrinkage formula divides all hours lost to internal and external shrinkage by total paid hours, then multiplies the result by 100.
Shrinkage % = (External Shrinkage Hours + Internal Shrinkage Hours) / Total Paid Hours x 100
Numerator and denominator must cover the same period and the same group of agents. Call Centre Helper notes that shrinkage is normally expressed over a full 12-month period, so a single month is best read as a snapshot.
Shrinkage and utilisation are two sides of the same figure, since shrinkage plus utilisation equals 100 percent. A 30 percent shrinkage rate therefore means 70 percent utilisation, and every call centre shrinkage calculation follows this structure.
How Do You Calculate Shrinkage in a Call Centre? A Worked Example
Consider a team of 100 agents on 40-hour weeks over a four-week operating month, which gives 16,000 paid hours. The hours below are illustrative assumptions, chosen to show how this call centre shrinkage calculation works.
| Component | Hours | Share of Paid Hours |
|---|---|---|
| Paid breaks | 2,000 | 12.5% |
| Unplanned absence | 1,280 | 8.0% |
| Training and coaching | 800 | 5.0% |
| Team meetings | 480 | 3.0% |
| System downtime | 160 | 1.0% |
| Total | 4,720 | 29.5% |
Shrinkage = 4,720 / 16,000 x 100 = 29.5 percent. That means about 70.5 of every 100 paid agents are available for customer contacts at any moment.
This snapshot leaves out annual leave and public holidays, which a full annual calculation must include. They added up to roughly 12 percent in Call Centre Helper’s annual worked example, so a complete agent shrinkage rate is usually higher than a four-week snapshot suggests.
How Do You Turn Shrinkage Into Staffing Numbers?
The required roster FTE formula converts productive demand into the number of agents that must actually be scheduled.
Required Roster FTEs = Required Productive FTEs / (1 – Shrinkage %)
If the forecast needs 100 agents on calls and shrinkage is 29.5 percent, the roster needs 100 / (1 – 0.295) = 142 agents. This is correct staffing, because only about 70.5 percent of a rostered team is available at any time.
Why Does Adding the Percentage Fall Short?
A common shortcut adds the percentage to the requirement, so 100 plus 29.5 percent gives 130 agents. But shrinkage applies to the rostered team, and 130 agents leave only 130 x 0.705, about 92, available. That is roughly 8 short. Call Centre Helper calls this the shrinkage trap.
In a spreadsheet, the same two calculations look like this.
Shrinkage % = (SUM(UnproductiveHours) / TotalPaidHours) * 100
Required roster FTEs = RequiredProductiveFTEs / (1 – (ShrinkagePercent / 100))
Why Does an Accurate Shrinkage Figure Matter?
Workforce management shrinkage feeds directly into rostering, so an error in either direction has a cost. The table below keeps the same 100-agent requirement and assumes a fully loaded cost of Rs 400 per agent hour over 2,080 paid hours a year, which is an illustrative assumption.
| Forecast vs Actual | Agents Rostered | Result |
|---|---|---|
| Forecast 35%, actual 30% | About 154 instead of 143 | About 11 extra agents, roughly Rs 91 lakh a year |
| Forecast accurate | About 143 | Service level met at the planned cost |
| Forecast 30%, actual 35% | About 143 instead of 154 | About 7 agents short, so longer queues and SLA risk |
Over-forecasting shows up as payroll, while under-forecasting shows up as service failures. Neither error is visible on a roster that looks full.
What Is a Good Shrinkage Percentage in a BPO?
Most contact centre professionals see shrinkage between 30 and 35 percent, and Dimension Data’s benchmarking report puts the average at 35 percent. The right agent shrinkage rate for a specific account still depends on shift pattern, leave policy, and process complexity.
A very low figure can also be a warning sign. If it comes from cutting training or coaching, it is likely a false economy, because error rates and repeat contacts tend to rise later.
In workforce management, shrinkage should also be read alongside occupancy and service level, since staffing calculations depend on all three.
Shrinkage is one floor metric among many. This guide to team leader KPIs shows how it sits next to handle time, adherence, and occupancy.
How Does Shrinkage Change Across the Day and Year?
Shrinkage is not flat. Call Centre Helper notes that it tends to peak between 9 and 11 am and again around 2 to 3 pm, when most team meetings are booked, and that it rises during holiday periods.
Intraday shrinkage explains why a 30 percent daily average can still leave a peak interval short. If 50 agents are rostered on a 30 percent assumption, 35 are available. A team huddle during a shift handover that pushes shrinkage to 45 percent leaves about 27.
Workforce management shrinkage reports usually show a daily average, which hides these swings. Reviewing intraday shrinkage by interval, and expressing it in hours per day or week as well as a percentage, gives a truer picture, especially where part-time staff make a one-hour meeting a much larger share of the day.
What Common Mistakes Distort the Shrinkage Figure?
Five mistakes account for most inaccurate shrinkage figures.
- Adding the percentage to the requirement instead of dividing by one minus shrinkage.
- Using different periods or agent groups for the numerator and the denominator.
- Leaving out annual leave, public holidays, or other rarely used leave types.
- Ignoring part-time staff, for whom a fixed meeting takes a larger share of paid time.
- Calculating internal shrinkage on all paid hours instead of on the hours agents are actually present.
The last point matters because coaching and meetings only take time from agents who are at work. Some planners therefore calculate external shrinkage first, then apply internal shrinkage to the hours that remain.
How Can a BPO Reduce Shrinkage Without Cutting Training?
Shrinkage splits into controllable and uncontrollable time, and control efforts work best on the controllable share.
- Move team meetings, huddles, and one-to-ones out of peak call intervals.
- Track absence and lateness by team, since a persistently high unplanned rate in one team often points to a management or morale issue.
- Fix recurring system problems, because every outage is unplanned internal shrinkage.
- Protect coaching and training time, since cutting it lowers the number while raising errors and repeat contacts.
Attrition adds to the load as well, since every leaver triggers recruiting and training hours. The BPO attrition formula shows how to measure that churn before planning around it.
How Does AI Change Shrinkage in a BPO?
AI voice agents change the calculation for the share of work they handle. They do not take breaks, leave, or coaching sessions, so that share of volume carries no agent shrinkage.
Platform downtime, retraining, and monitoring still need planning. Calls that need judgement also still go to human agents, whose rosters carry normal shrinkage.
Suppose AI resolves 30 percent of a workload that needs 100 productive agents. Human demand falls to 70 productive agents, and at 29.5 percent shrinkage the roster needs 70 / 0.705, about 99 agents instead of 142. That is roughly 43 fewer rostered agents for the same workload.
DialNexa’s India voice AI report covering more than a million AI-assisted calls recorded an 89 percent goal completion rate on inbound calls, with median AI response latency under one second. That points to routine inbound work being completed without a human on the line.
Teams adding AI to the mix should track it with its own measures. This guide to voice agent KPIs covers how AI agent performance is measured.
Conclusion
The BPO shrinkage formula divides internal and external shrinkage hours by total paid hours and multiplies by 100. Most contact centres land between 30 and 35 percent, and the figure becomes useful once the required roster FTE formula turns it into a roster.
Accuracy depends on a consistent period, complete inputs that include leave, and dividing by one minus shrinkage instead of adding a percentage. Each of those small choices decides whether a full roster actually covers the forecast.
Teams weighing how AI could reduce routine call volume, and the roster that volume requires, can test DialNexa on real call scripts before changing any staffing plan.
FAQs
1. What is the BPO shrinkage formula?
Shrinkage % = (external shrinkage hours + internal shrinkage hours) divided by total paid hours, multiplied by 100. External shrinkage covers leave, sickness, lateness, and paid breaks, while internal shrinkage covers meetings, coaching, training, and system downtime. Both must use the same period and agent group.
2. How do you calculate shrinkage in a call centre with an example?
With 100 agents on 40-hour weeks over four weeks, paid hours are 16,000. If breaks, absence, training, meetings, and downtime total 4,720 hours, shrinkage is 4,720 divided by 16,000, or 29.5 percent. About 70.5 of every 100 paid agents are available.
3. What is a good shrinkage percentage in a BPO?
Most contact centre professionals see shrinkage between 30 and 35 percent, and Dimension Data’s benchmarking report gives an average of 35 percent. The right target depends on shift pattern, leave policy, and process, and a very low figure may mean training is being cut.
4. How do you convert shrinkage into required staffing?
Divide required productive agents by one minus shrinkage. At 29.5 percent shrinkage, 100 productive agents need 100 divided by 0.705, about 142 rostered. Adding 29.5 percent to 100 gives only 130 agents, which leaves roughly 8 seats uncovered at any moment.
5. What is the difference between planned and unplanned shrinkage?
Planned shrinkage covers scheduled time such as leave, training, coaching, and meetings, which can be moved or reduced. Unplanned shrinkage covers events such as sickness, lateness, and system downtime, which cannot be scheduled and need buffers and monitoring instead of rescheduling.
6. Can AI voice agents reduce shrinkage in a BPO?
AI voice agents take no breaks, leave, or coaching time, so the volume they handle carries no agent shrinkage. Human rosters still need shrinkage planning for the remaining calls, and platform downtime, retraining, and monitoring need separate planning of their own.
