Erlang C, named after the Danish engineer Agner Krarup Erlang, gives the probability that a caller will have to wait in a queue, based on how many calls arrive, how long they last and how many agents are available. Workforce planners usually run it in reverse: they set a service level target and look for the smallest number of agents that meets it.
Inputs:
- Call volume: calls expected per interval, for example per 30 minutes.
- Handle time: the average handle time, meaning talk time plus hold time plus wrap-up.
- Target: for example 80% of calls answered within 20 seconds.
The output is the number of agents needed, with the expected average speed of answer and agent occupancy at that staffing level.
Erlang C assumes that callers never hang up while waiting and that every agent can take every call. Real callers do abandon, so when waits are long the formula tends to overestimate the agents needed; variants such as Erlang A account for abandons. Its result is also the number of agents on the phones at the same time. To get the number of people to schedule, planners add shrinkage: breaks, training, absence and other time away from the phones.