# Forecast Accuracy Targets

Topic Views - 3370

Currently, we have 3 forecast accuracy targets. For intervals with very low volume which are from 0000H-0730H interval, our target is 60%. For intervals with high-volume which are from 0800H-1730H, target is 80%, and from 1800H-2330H interval which have mid-volume, target is 70%. Low-volume intervals have 0-5 calls, for high-volume intervals we have over a hundred calls, and for mid-volume it is somewhere in between. What can you say with this set-up? Is our forecast accuracy target okay?

To add, we measure our forecast accuracy per interval and per day. Wherein our target per interval is 80% which means, we need to hit the per interval targets 80% of the time. For the daily it's 90%. We compute it as the overall volume of all the intervals versus the overall forecast for the day.

We follow the formula 1-ABS(Forecast-Actual)/Actual * 100

Are the parameters too strict? Or too linient? Or just okay?

Helper

Call Centre Helper

The targets sound about right I would say. Some organisations can get much better than this but it depends upon a number of factors.

- How stable the call volumes are

- How many years of data do you have

- How well built are the links with other departments

- How much can you model changes in demand (e.g Weather, Special events)

Did you watch the webinar that we did on this topic last week?

Thanks for the response Jonty. Oh, I miss the webinar. 'Wish I could have a peek on it. Here are my answers":

How stable the call volumes are - very, very unstable. we observed the "randomness" in the call arrival.

How many years of data do you have - I have 3 years data. But when forecasting, we only consider the last 4 weeks data.

How well built are the links with other departments - What do you mean by this question?

How much can you model changes in demand (e.g Weather, Special events) - What do you mean by this question?

Thanks.

Helper

Call Centre Helper

Here is a link to the webinar