A formula that gives the chance a contact has to wait, from the workload and the number of agents. Contact centres use it to find the fewest agents that still meet a service level, like 80% of calls answered within 20 seconds.

Paid time agents can't take contacts: breaks, meetings, training and absence. The calculator adds it on top of the agents needed at any moment.

Does it work for chat?

Erlang C assumes one contact per agent at a time. For chat, divide the result by how many chats an agent handles at once, as a rough guide.

Yes, with no sign-up. It runs in your browser, so nothing you type is sent to us.

Erlang C staffing calculator

Enter how many contacts you get, how long each takes and your service level target, and get the number of agents to schedule, worked out with the Erlang C formula.

Contacts in the period

Period length (minutes)

Average handle time (seconds)

Talk or chat time, plus wrap-up.

Shrinkage (%)

Breaks, training, absence.

Service level target (%)

Answered within (seconds)

Agents to schedule

14 agents answering at any moment, plus shrinkage. Expect 86.7% answered within 20 seconds, an average wait of 13 seconds and 71.4% occupancy.

Workload: 10 erlangs (contacts × handle time ÷ period). Calculated with the Erlang C formula, which assumes contacts arrive at random and wait in one queue. For chat, divide by how many chats an agent handles at once.

Erlang C staffing examples

Worked examples with the same formula as the calculator above.

100 calls in 30 minutes

Average handle time 3 minutes (180 seconds). Target: 80% answered within 20 seconds. Workload = 100 × 180 ÷ 1,800 = 10 erlangs 13 agents answer 79.6% within 20 seconds, just short 14 agents answer 88.8%, so 14 are needed With 30% shrinkage: 14 ÷ 0.7 = 20 agents to schedule

Why one agent matters so much

Same workload of 10 erlangs: 11 agents: 39.0% answered within 20 seconds 12 agents: 64.0% 13 agents: 79.6% 14 agents: 88.8% Near the target, each extra agent moves the service level a long way.

Chat

120 chats an hour, 6 minutes each, and agents handle 2 chats at once. Work it out as if each agent took one chat at a time, then divide by 2. Treat the result as a guide: Erlang C assumes one contact per agent.

Let unclench.ai answer your customers from your help content, around the clock.

Try it on your website Start your 7-day free trial

Questions, answered.

What is Erlang C?: A formula that gives the chance a contact has to wait, from the workload and the number of agents. Contact centres use it to find the fewest agents that still meet a service level, like 80% of calls answered within 20 seconds.
What is shrinkage?: Paid time agents can't take contacts: breaks, meetings, training and absence. The calculator adds it on top of the agents needed at any moment.
Does it work for chat?: Erlang C assumes one contact per agent at a time. For chat, divide the result by how many chats an agent handles at once, as a rough guide.
Is it free?: Yes, with no sign-up. It runs in your browser, so nothing you type is sent to us.

Fewer repeat questions means fewer agents.

unclench.ai answers the repeat questions from your help content, so your team's time goes to the contacts that need a person.