Agents saved by pooling 3 queues

Assumptions

• 3 queues of equal size, 300-second handle time, a 30-minute interval, 80% answered within 20 seconds, no occupancy ceiling and nobody hanging up.
• The pooled figure assumes every agent can take every contact at the same handle time, which is the ceiling perfect cross-training would reach rather than a promise.
• Erlang C throughout, so no allowance for abandonment; a queue where people hang up needs fewer agents than shown, and the saving from pooling is smaller.

Change any assumption in the queue pooling calculator

Frequently asked questions

Why do 3 small queues cost so much more than one big one?

Every queue needs agents above its own workload to keep waits short, and that slack is proportionally larger the smaller the queue. Separating the work buys that slack once per queue. The slack is doing the same job each time, absorbing random arrivals, so most of it is wasted duplication.

Can a real centre actually capture this saving?

Rarely all of it. This assumes anyone can take anything at the same speed. In practice only some agents hold a second skill, that work usually takes them a little longer, and which queues you join together matters as much as how many people you train. The cross-training simulator on this site estimates how much of the saving a specific skill design reaches.

The queues are not the same size.

Open the pooling calculator from any row and enter the real volumes. Unequal queues still save agents, and the biggest gains come from the smallest queues joining the pool. Watch the handle times too: pooling a queue whose calls are several times longer than the rest can make the short queue worse off.