Forecasting In 15 Minute Intervals

Forecasting

We forecast our call volumes in 15 minute intervals. Who’s to say the calls will come in evenly throughout that 15 minute period. We could get 50% of the calls in the first 5 minutes.

How do other people get round this in their contact centers?

Question asked by pete

If you are using Erlang

If you are using Erlang to get your staffing requirements, or a simulation tool, then the bunching will be accounted for within both approaches (the latter should be more accurate with sufficient time, the former easy to use in Excel sheets for ease of understanding).

Call arrival is random, but based on patterns, so the interval you use should depend on the data available to you and the reasons for planning (if you book down to 15 min activities a lot, then it’s a good interval size to use for example). There is nothing to stop you going smaller, but you gain very little benefit usually, and in most cases, getting the historical and real time data at lower intervals is usually difficult. Some places will use 30 minute intervals and it works equally as well, and in you can even push out to an hour (or longer, but you begin to lose the granualrity to plan for and accuracy in calculation).

With thanks to robtuck

Calls will be skewed around the top of the hour

It is quite likely that calls will be skewed around the top of the hour

See point 7 in this article – 7 Ways to Improve the Accuracy of Your Contact Centre Forecasts

I agree with Rob that going less than 15 minutes may not gain you more. The problem is that the lower time increment that you use, the lower that Erlang based calculations can be in accuracy.

With thanks to Jonty

Monet

Does anyone use Monet? If so what intervals does it report at/record call data at from the box.

With thanks to nbrady

15 Minute Intervals

I found this online

Monet Features

Set up business rules and center group parameters, select 60, 30, or 15 minute time increments to build and collect current or historical data, select center hours of operation, and initialize center service level metrics.

With thanks to Jonty

Author: Jonty Pearce