Capacity Planning: Definition, Formula, and a Worked WFM Example
Capacity planning, in workforce management, is the process of converting a forecast of work volume into the number of staffed hours and full-time equivalents needed to meet a service target, after accounting for handle time, occupancy, and shrinkage.
Capacity planning is the workforce management discipline that answers one question: how many people do we need, and when? It starts with a forecast of demand, such as calls, chats, tickets, or cases, converts that demand into hours of work using handle time, and then converts hours of work into staffed hours and full-time equivalents by accounting for the time agents spend not handling contacts. The output is a hiring plan and a schedule, and its accuracy is what separates a center that meets service level with a stable team from one that lurches between overtime and idle agents.
What Is Capacity Planning?
The term is borrowed from manufacturing and IT, where it means matching plant or server capacity to demand. In a contact center or any queue-driven operation, capacity is agent time, and planning it means projecting workload months ahead so that hiring, training, and scheduling all land before the demand does. A new agent hired today is productive in six to ten weeks, so a plan that discovers a shortfall in the current month has already failed.
Capacity planning sits between forecasting and scheduling. Forecasting predicts the work; scheduling assigns named people to intervals. Capacity planning converts the forecast into the headcount the schedule can draw on. When the three are run by separate people, the handoffs are where the errors hide, which is why forecast accuracy and shrinkage each get measured on their own.
The Formula Chain
Capacity planning is a chain of conversions, each with its own assumption.
- Workload hours = Forecast contacts × Average handle time (in hours)
- Base staff hours = Workload hours ÷ Target occupancy
- Staffed hours required = Base staff hours ÷ (1 − Shrinkage)
- FTEs required = Staffed hours required ÷ Paid hours per FTE for the period
Step 2 accounts for the fact that agents cannot be handling contacts every second they are logged in; at 85 percent occupancy, every hour of workload needs 1.18 hours of logged-in time. Step 3 accounts for the paid time agents are not available at all: breaks, meetings, training, absence, and leave. Step 4 turns hours into people.
For inbound voice, step 2 is usually replaced by an Erlang C calculation that derives the agents needed per interval directly from volume, handle time, and a service level target, which captures the effect of random arrivals better than a flat occupancy assumption. The site's Erlang C calculator does that arithmetic; the chain above is the month-level version that a hiring plan uses.
Worked example: a 50-agent inbound team planning for next quarter
An inbound support team forecasts 62,000 calls for October. Average handle time is 6.5 minutes. The team targets 85 percent occupancy and runs 30 percent shrinkage. A full-time agent is paid for 173.3 hours in a month (2,080 ÷ 12).
| Step | Calculation | Result |
|---|---|---|
| Workload hours | 62,000 × 6.5 ÷ 60 | 6,716.7 hours |
| Base staff hours at 85 percent occupancy | 6,716.7 ÷ 0.85 | 7,902.0 hours |
| Staffed hours at 30 percent shrinkage | 7,902.0 ÷ 0.70 | 11,288.6 hours |
| FTEs required | 11,288.6 ÷ 173.3 | 65.1 FTEs |
The team has 50 agents. The plan says it needs 65 for October, a gap of 15 FTEs. With a six-week hire-to-productive cycle, requisitions must open by mid-August. If the team instead discovers the gap in the first week of October, its only options are overtime, which pushes occupancy past 85 percent and raises attrition, or missing service level.
The sensitivity is worth seeing. Cutting shrinkage from 30 to 25 percent reduces the requirement to 60.8 FTEs; a 5 percent forecast error in either direction moves it by about 3.3 FTEs. That is why capacity planners spend most of their time on shrinkage and forecast accuracy rather than the arithmetic.
Requirements are computed in hours because that is what the workload is. They are converted to FTEs only at the end, because that is what a requisition is. Converting too early, by planning in "agents," hides the effect of part-time schedules, overtime, and shrinkage changes.
Long-Term vs Short-Term Capacity Planning
| Long-term plan | Short-term plan | |
|---|---|---|
| Horizon | 3 to 18 months | 1 to 6 weeks |
| Unit | FTEs by month | Staffed hours by interval |
| Main decisions | Hiring classes, attrition replacement, outsourcing | Overtime, voluntary time off, shift bids |
| Key assumptions | Attrition rate, hire-to-productive time, annual shrinkage | Interval-level forecast, weekly shrinkage, adherence |
| Tools | Spreadsheet capacity model, FTE calculator | Erlang C, WFM scheduling software |
The long-term plan adds attrition to the chain: if the team loses 4 percent of agents a month, a 65-FTE requirement in October needs hiring in August for October's demand plus replacements for the agents who will leave between now and then. The attrition rate entry covers that calculation.
Common Mistakes
- Planning to average volume. Demand arrives in peaks. A plan built on the monthly average staffs the average interval and fails every busy one. Plan to the interval-level forecast and aggregate up.
- Using last year's shrinkage. Shrinkage moves with the calendar (holidays, training pushes) and with attrition (new hires have more training time). Use a forecast of shrinkage, not a historical constant.
- Ignoring hire-to-productive time. A hire counts as capacity when they take contacts at full handle time, not when they sign an offer. New agents also inflate handle time for their first months.
- Treating occupancy as a lever. Raising the occupancy target from 85 to 92 percent "finds" 7 percent capacity on paper and loses it back in attrition and burnout within two quarters.
Capacity Planning in Contact Centers and Remote Teams
Remote and outsourced teams change one input more than the others: shrinkage. A distributed team's shrinkage is harder to observe, because breaks, meetings, and unplanned absence are not visible on a floor, and it tends to be underestimated in the plan and discovered in the interval data. A BPO running a 50-agent remote team for a client planned at 25 percent shrinkage and measured 33 percent once it compared scheduled hours against logged-in hours. The 8-point gap was worth about 7 FTEs, which was the whole of the service-level miss the client had been escalating.
For non-queue work, such as back-office processing or a remote development team, the same chain applies with cases or tickets in place of calls and cycle time in place of handle time. The occupancy step is replaced by a utilization target, and the output is still staffed hours.
How to Track the Inputs
Capacity planning is only as good as its shrinkage and handle-time inputs, and shrinkage in particular comes from time and attendance data. HiveDesk records clock-in, clock-out, and break times from desktop, mobile, and browser apps, so scheduled hours can be compared against hours actually worked to measure shrinkage rather than assume it. Attendance records give unplanned absence by day, leave management gives planned absence, and timesheets give paid hours by employee, which is the denominator of the FTE calculation. HiveDesk does not forecast contact volume or compute Erlang C staffing; it supplies the hours-and-attendance side of the plan. The plan is $5/user/month with a 14-day free trial and no credit card required. The FTE calculator and the call center capacity planner handle the conversions.
Measured Shrinkage, Not Assumed
HiveDesk's time, attendance, and leave records show the paid hours your team actually worked, so capacity plans start from real shrinkage. $5/user/month, 14-day free trial.
Related Terms
- Shrinkage: the largest and least certain input
- Forecast accuracy: how good the volume input is
- Occupancy rate: the productive-time assumption in step 2
- Average handle time: the conversion from contacts to workload hours
- Attrition rate: the replacement hiring the long-term plan adds
- Workforce management: the discipline capacity planning belongs to
Frequently Asked Questions
What is capacity planning in a call center?
Converting a forecast of contacts into the staffed hours and FTEs needed to meet a service target, by applying handle time, an occupancy or Erlang C model, and shrinkage. The output drives hiring and scheduling.
What is the capacity planning formula?
Workload hours = contacts × handle time; staff hours = workload ÷ occupancy; staffed hours = staff hours ÷ (1 − shrinkage); FTEs = staffed hours ÷ paid hours per FTE. For inbound voice, Erlang C replaces the occupancy step at interval level.
What is the difference between capacity planning and forecasting?
Forecasting predicts the volume of work. Capacity planning converts that volume into the people and hours needed to handle it. Scheduling then assigns named people to intervals.
How far ahead should you capacity plan?
Far enough to hire: at least the hire-to-productive cycle plus recruiting time, typically 3 to 6 months, with a rolling 12- to 18-month view for budgeting.
Why does shrinkage matter so much in capacity planning?
Because it is applied as a divisor. At 30 percent shrinkage, every hour of required staff time needs 1.43 paid hours. A 5-point error in shrinkage moves a 65-FTE requirement by 4 to 5 FTEs.
Can you do capacity planning without WFM software?
Yes, at the monthly level, with a spreadsheet and the formula chain above. Interval-level planning for inbound voice needs an Erlang C calculation, which a calculator can provide.
Browse more workforce management terms in the glossary.