Some call centres have a reporting bias towards the percentage of calls answered/abandoned, some the percentage of calls answered within a given target time, some the average wait time, and others the longest wait.
We will assume that we are aiming to answer 80% of calls within 30 seconds, and that we have access to a spreadsheet function of the form
Agents(80%, 30, calls, aht)
that will tell us the number of advisors required to deliver that service level with that number of calls in an hour, answering at that AHT. (If you do not have such a function in your spreadsheet program, you may be able to download and install an add-in from the Internet.)
We know that Erlang calculators try to estimate a level of availability we need to build in to our models to simulate the impact of random call arrival. Such models are typically very helpful on an hour-by-hour basis, but how can we use them to estimate staffing required for a week?
Follow the links for more.