A practical guide to forecasting demand, calculating how many agents you need and building schedules that hit service targets without burning people out.
Calls, chats and emails arrive in uneven waves through the day and week. Too few agents means long waits and abandoned calls; too many means paying people to sit idle. Unlike many shift businesses, call centers plan staffing in 15- or 30-minute intervals and measure results against a service level, such as a target percentage of calls answered within a set number of seconds. Our call center management guide covers the wider operation.
| Term | Meaning |
|---|---|
| Contact volume | Number of calls, chats or emails in an interval |
| Average handle time (AHT) | Average talk time plus hold time plus after-call work |
| Workload | Volume times AHT, expressed in hours |
| Service level | Share of contacts answered within a target time, for example 80% within 20 seconds |
| Occupancy | Share of logged-in time agents spend handling contacts |
| Shrinkage | Paid time when agents are not available to take contacts: breaks, training, meetings, absences |
| Schedule adherence | How closely agents follow their scheduled activities |
Workload is the starting point: if 120 calls arrive in an hour and AHT is 6 minutes, the workload is 720 minutes, or 12 agent-hours. But you cannot staff exactly 12 agents and still meet a service level, because calls arrive randomly and some will queue. Planners commonly use the Erlang C formula, built into most workforce management tools and available in online calculators, to work out how many agents are needed for a given volume, AHT and service level target. The answer is always higher than raw workload, and the gap is larger for small teams.
| Input | Value |
|---|---|
| Calls | 120 |
| AHT | 6 minutes |
| Raw workload | 12 agent-hours |
| Service level target | 80% answered in 20 seconds |
| Agents needed on the phones (from an Erlang-style calculation) | More than 12; the exact number depends on the model and inputs |
These figures are for illustration only. Use a calculator or your workforce management tool with your own data.
Agents are never available for 100% of their paid time. Breaks, meals, coaching, team meetings, training, system problems, vacations and sick days all reduce available time. If shrinkage is 30%, you need to schedule the required agents divided by 0.7 to have enough on the phones. Measure your own shrinkage by category and update it regularly; it varies a lot between centers.
| Category | Planned or unplanned |
|---|---|
| Breaks and meals | Planned |
| Coaching and team meetings | Planned |
| Training | Planned |
| Vacation and holidays | Planned |
| Sick leave and call-outs | Unplanned |
| Late arrivals and early departures | Unplanned |
Breaks must be planned so queues stay covered. Stagger breaks across the team and avoid putting them in peak intervals. Under federal law, short rest breaks (usually 20 minutes or less) are generally paid, while bona fide meal periods of 30 minutes or more when the agent is fully relieved can be unpaid. Several states require meal and rest breaks at specific times, such as California; see our California break rules guide. Time agents spend logging in, starting up computers and opening required systems before taking calls is generally working time and must be paid.
Adherence measures whether agents are doing the scheduled activity at the scheduled time, such as being available for calls rather than on break. Conformance compares total time worked with scheduled time. Track both, but use them to support agents, not punish them for things outside their control, like a long call that runs into a break. Clear break rules and a quick way to request schedule changes improve adherence more than strict policing.
Agent turnover is costly in call centers because of training time. Our guide to call center agent motivation covers ways to keep good agents.
Reviewing forecast accuracy, service level and overtime together each week helps catch these problems early.
Many call centers now include agents working from home. The same scheduling principles apply, but you also need reliable time tracking that covers login, breaks and logout from any location, and clear rules for paid start-up time, technical problems and internet outages. Make sure remote agents get the same access to schedules, swaps and team communication as on-site agents.
EZTime helps call centers build and manage shift schedules. Managers use scheduling software to create staggered shifts, publish them early and handle swaps; agents clock in and out from the office or home with the attendance clock or mobile app; and actual hours feed payroll. See EZTime for call centers. Erlang-style staffing calculations are typically done in a dedicated forecasting tool or calculator before shifts are built.
Forecast volume and handle time by interval, use an Erlang-style calculation to meet your service level, then add shrinkage.
Paid time when agents are not available to handle contacts, such as breaks, training, meetings and absences.
Many centers use targets like 80% of calls answered within 20 seconds, but the right target depends on your customers and business.
How closely agents follow their scheduled activities, such as being available for calls rather than on break at the scheduled time.
Time spent starting computers and opening required systems before taking calls is generally working time for non-exempt agents.
Two weeks is a common target, and some locations have predictive scheduling rules.
Part-time shifts are often an efficient way to cover predictable peaks, such as mid-morning or early evening, without overstaffing quiet hours.
EZTime schedules include shift reminders and real-time labor costing, and the clock can refuse punches outside an assigned shift.