Last Updated August 2016
This week we will be looking at what to consider for short term forecasting and how to avoid some common pitfalls.
What is Short Term Forecasting?
By short term forecasting I am referring to the forecasts that you will create requirements and apply the schedules against. Or in simple terms, the ones you will use for intra-day management.
In terms of the forecast volumes for each contact type it is important to consider the historical data. If you are using a Workforce Management solution, you will have the benefit of the history so will be able to either let the application forecast for you or manually override.
It is also worth considering whether you would like to exclude any extreme events that may skew the historical profiles. For example, did you have a day where the volumes were impacted by a system issue? This may have caused spikes in contact volumes that need to be excluded or smoothed.
You may also want to look at the call profiles on a weekly basis, i.e. do volumes increase or decrease at certain points in the month. In many financial organisations, the monthly billing cycle may cause uplift in contacts or in a collections environment, mid month may see lower incoming call volumes.
Weighting of Forecasting Model
The other major factor to consider is the how the forecasting model is weighted. This means how much weighting is applied to the previous x weeks of volumes. You may want to equally weight each previous week if you work in a stable environment or you may want the last three weeks to influence the forecast. In some cases you may only want to look at a few weeks of historical information. This will all depend on the business and contact type however the key thing is to ensure the model is reviewed and updated on a regular basis. It is also worth tracking forecast accuracy over time using techniques such as MAPE (mean average percent error). This will ensure over and under forecasts do not compensate each other.
The Key to a Good Forecast
Whether you are using a WFM application or not, the key to a good forecast is:
- an in-depth knowledge of events and activities
- awareness of the daily, weekly and monthly trends
- and the ability to pull on historical or special event profiles.
If you found this blog post interesting, check out our Workforce Management Advice Hub for the latest information on WFM technology.