Support Staffing Planner · guide
Short answer: enough that the share of contacts answered within your target time is met in every half hour, plus the people who are paid but not on the queue. Dividing workload by a comfortable occupancy looks reasonable and misses the target badly at small volumes. The queueing formula that gets it right is Erlang C.
Between 09:00 and 09:30 you expect 30 contacts and each takes 300 seconds (five minutes) to handle. The workload is 30 × 300 / 1,800 = 5 erlangs: five agents would be busy the whole half hour if nothing ever queued.
The occupancy-only rule would have staffed 6 and missed the target by a wide margin.
Once every half hour has a requirement, you need shifts that cover all of them for the least paid time. The Support Staffing Planner takes your contacts per interval and the shift lengths you allow, finds the cheapest combination of shifts that covers every interval (an exact solution, not a rule of thumb), and compares it with flat peak staffing, with full-length shifts only, and with the occupancy rule. It also prices each service target and shows what a shorter handle time or fewer contacts would save.
Erlang C assumes random arrivals and that nobody hangs up while waiting, so real service is usually slightly worse than the formula says. Treat the result as a planning aid and check it against a few real weeks.
The common phone target is 80% answered within 20 seconds, but it is a choice, not a rule. Email and chat teams often use 90% within a few minutes. The planner prints what each target costs so you can see what the last ten points are worth.
Because queues are not smooth. At 5 erlangs, 6 agents is 83% occupancy, which sounds reasonable, but only about 45% of contacts are answered within 20 seconds. Eight agents answer about 86%.
Add up the paid time agents are not on the queue: breaks, meetings, training, leave and sickness. Thirty percent is a common starting figure for phone teams; use your own payroll and schedule data.