Create a long term forecast
Long Term Forecasts employ industry standard algorithms to generate forecasts for queues over periods of up to 36 months. Long Term Forecasts are generated using one of the available algorithms and at least 12 months of historical queue Entity that represents demand in WFM. Queues help predict workload by multiplying the volume of customer interactions by their expected handling time. data.
The algorithms used by Long Term Forecasts are:
-
Prophet: an open-source algorithm released by Meta, used for forecasting time series data based on an additive model where non-linear trends are fit with yearly, weekly, and daily seasonality, plus holiday effects.
-
Strategic Forecasting: Verint’s proprietary Strategic Forecasting algorithm uses an enhanced forecasting method based on historical data, using advanced regression logic, to create long-term One to five words that are meaningful to a specific type of business, or phrases that stand out in interactions in Speech and Text Analytics. forecasts at the daily level of granularity that are based on past trends, seasonality, and past and future special events.
Once active, the forecast generated, feeds into any scheduling periods that contain the same queues and any date ranges that overlap with the program range. A weighting value determines how much of an effect that the long term forecast data has on the tactical forecast data.
Before you begin
Create a program with at least one program range and queue.
Procedure
-
Go to Forecasting and Scheduling. Under Forecasting, select Long Term Forecast Feature in WFM used to estimate the required resources for a specific scheduling period based on historical data. The data in a forecast includes interaction and work volume, and Average Handling Time (AHT)..
-
Select Programs in the left pane, then select the program to use from the list.
-
Select Program Ranges in the left pane, then select from the available program ranges.
-
Select New Forecast from the upper right of the page.
-
Add a Name and Description for the forecast.
-
For Algorithm Selection, choose Automatic or Manual to determine how one of the following algorithms is selected for the forecast.
-
Prophet
-
Strategic Forecasting
If Automatic, one of the algorithms is selected based on how closely the forecast it generates matches the past performance of your queue using the available historical data. For Automatic, we recommend that you have at least 12 months data plus the length of the forecast period, but up to a maximum of 24 months in total. For Manual, you must have 12 months of historical data, for all forecast period lengths.
Example:
The amount of historical data required for each of these periods:
6 months: 18 months for Automatic, and 12 months for Manual.
18 months: 24 months for Automatic, and 12 months for Manual.
-
-
Select the link to see the details of any Special Events associated with the queue for the forecast period. By default, all special events are enabled, cancel the selection of any you do not want to be included in the forecast.
-
Select Create Forecast.
While the forecast is generating, it appears in the left side pane with a spinner next to it, indicating that the forecast is in progress. At this stage, you cannot select it. Once the forecast is complete, the spinner is no longer visible, and you are able to select the forecast and view the data.
If there is not enough historical data for one or more of the queues, the forecast is still generated, but is zero for those queues with insufficient data. An indication is shown on the forecast, and also on each of the queues where there was insufficient data.
-
Optional: To use the forecasted data in tactical forecasting, select Set as Active.
The data can be used in tactical forecasting when the dates of the program range overlap the dates of the scheduling period for the queue.