Erlang A calculator (with abandonment)

Erlang C assumes callers wait forever. Erlang A adds caller patience — modelled as average time before someone hangs up — and predicts how many will abandon if the queue gets long.

Inputs

Contacts in interval

Interval length (minutes)

Average handle time (seconds)

Average caller patience (seconds)

Mean time before a queued caller hangs up. 60–120s is typical for voice; chat patience is usually higher.

Max acceptable abandonment (%)

Target ASA (seconds) — optional

Shrinkage (%) — optional

Results

Agents required (net)—

Scheduled (after shrinkage)—

Predicted abandonment—

Mean wait (all callers)—

Probability of waiting—

Occupancy—

Offered load (Erlangs)—

Sensitivity

Agents	Abandon %	Mean wait (s)	Occ %

Erlang A is more realistic than Erlang C when queues are long and callers do hang up. Patience is typically distributed exponentially in the model; in real life it’s more complex (some callers abandon in the first 5 seconds, others wait minutes). Treat the abandonment number as an indicator rather than an exact prediction.

How it works

Erlang A extends Erlang C with one crucial addition: caller patience. Real callers don't wait forever — some abandon when the queue gets long, which both relieves the queue and represents lost service. By modelling that patience (entered as an average number of seconds before a caller hangs up) Erlang A predicts your abandonment rate alongside service level, and usually returns a slightly lower staffing requirement than Erlang C for the same target, because abandonment shortens the queue. It's the right tool when abandonment is a meaningful share of your contacts.

Common questions

What is caller patience and how do I estimate it?

Patience is the average time a caller will wait before abandoning — often 60–120 seconds for voice. Estimate it from your abandonment data, but treat it as an indicator: you only ever observe the patience of callers who actually abandoned, which understates everyone who stayed.

Why does Erlang A need fewer agents than Erlang C?

Because some callers abandon, the queue is shorter than Erlang C's 'everyone waits forever' assumption implies, so fewer agents hit the same service level. Lean on that too hard, though, and you're effectively staffing on the back of customers giving up.