What method should I choose when I create a short-term forecast?
Workforce management has four creation method options when creating a short-term forecast. Each option has specific uses, capabilities, and requirements.
|Automatic Best Method||
This AI powered forecasting method is the most sophisticated methodology offered in workforce management. It includes:
Plus, if a custom model based on multiple methodologies that are equally weighted produces a better result than a single model, the forecast is classified as ensemble. You can configure this creation method to create forecasts from one to six weeks in length. For more information, see ensemble forecasting.
|This method requires historical data in Genesys Cloud. While Automatic Best Method works with as little as one week of historical data, it is much more powerful with more historical data. Specifically, to determine daily arrival patterns for scheduling, this methodology looks at up to 90 days of historical data. For seasonality detection, a full season or multiple seasons of data produce the best results.|
|Weighted Historical Index||This method allows the forecaster to weight, at a daily level, one or more weeks of historical data. It uses the resulting weighted average, by day, to output a forecast one week at a time.||
This method requires historical data in Genesys Cloud. However, it does give the option to import data from outside the platform. Because of the flexibility of this creation method, customers often use it when a lot of manual forecast manipulation is necessary.
|Weighted Historical Index with Source Data Import||This method is the same as Weighted Historical Index except that it asks you to import data from a file instead of using data in the Genesys Cloud platform.||
This method requires a data file with at least one historical week of data in it to use for the weight averaging process. New Genesys Cloud customers who do not have historical data on the platform often use this method.
This method is for customers who:
|This method requires a file containing a forecast that meets the file specifications. It does not require historical data.|