Forecasting
Use the Workforce Management (WFM) Forecasting tool to predict contact-center workload and staffing requirements, based on historical data or user-defined templates. WFM provides multiple methods of forecasting the workload and staffing requirements for work activities in the following sections:
- Using Historical Data
- Using Forecasting Events
- Setting Service Objectives
- Flexible Forecasting
- Deferred-Work Forecasting
- Multi-Skill Support
Start by creating one or multiple forecast scenarios. Creating multiple scenarios enables you to see the effects of changes to forecasting parameters, such as service objectives and predicted interaction volume. When you have determined the most satisfactory forecast, publish it, making it the Master Forecast on which schedule scenarios, and eventually the Master Schedule, are based.
If you choose to, you can derive workload forecasts from historical information that is either collected automatically by WFM from the Genesys system or imported from .csv files using the WFM Configuration Utility. You can also create workload and staffing forecasts as reusable templates. Once you have generated a workload prediction, WFM determines the staffing requirements needed to service the workload, taking into account any applicable service objectives.
Using Historical Data
WFM automatically collects historical data from Stat Server for all work activities handled by the Genesys platform encompassing all media, contact segments, and service types. Using Genesys Stat Server, rather than automatic call distribution (ACD) reports, provides you with far greater flexibility in defining and gathering statistics that provide an appropriate measure of contact center performance over time.
WFM analyzes interaction volumes and average handling time (AHT) in order to predict future trends for each work activity. This data enables WFM to build accurate forecasts for the anticipated workload, and to calculate the staffing required to meet that workload.
Using the WFM API, you can also develop a custom application that will enable Interaction Volume and AHT data from a third-party system to be imported directly into the WFM database. This is useful if you want to use WFM to forecast and schedule a type of work that is not being routed by Genesys.
Using Forecasting Events
WFM can track events that might affect interaction volume. A sales promotion or marketing campaign, for example, might cause a predictable peak in interaction volume. These types of events are entered in WFM Web for Supervisors and used by the advanced WFM algorithms. If an event recurs, the forecasting algorithms learn the impact of that event and account for its impact in future forecasts.
Setting Service Objectives
With WFM forecasting, you can set specific service objectives. You can also adjust these objectives and then rebuild the forecast, which provides a detailed "what-if" analysis of the potential impact of staffing or service-objective changes. WFM forecasting uses parameters to determine effects of different service objective settings, such as: *Interaction volumes
- Average handling time (AHT)
- Average speed of answer (ASA)
- Desired percentage of interactions handled within a target time (service level)
- Occupancy
- Maximum percentage of abandoned interactions.
Flexible Forecasting
WFM supports an unlimited number of forecasting scenarios, enabling you to create multiple forecasts and evaluate how changes in the parameters or the forecasting method that you use, affect expected service objectives. Resource planners can then easily create reliable forecasts, fine-tuning the results in tabular and graphical data views. You can also save forecast workforce data as templates for use in subsequent forecast building.
WFM offers several different forecasting methods of varying complexity:
- Template-Based—Good for work activities with little historical information or for activities with very predictable interaction traffic.
- Expert Average Engine—Good for work activities that have a reasonable amount of historical data or those that fluctuate more dramatically because of unknown factors.
- Universal Modeling Engine—Good for work activities with more than one year of historical data and accurate forecasting event information.
- Copy Historical Data—Good for work activities when you have some historical data, but not enough to use the Expert Average Engine or the Universal Modeling Engine. You can combine the historical data with overlap templates, which fill in gaps in the historical data.
- Use Value—Good for work activities if your site activity load is very regular. Applies a specific interaction volume or AHT to each time interval in the scenario.
Deferred-Work Forecasting
WFM is designed to consider deferrable work activities, such as e-mail, as inherently different from immediate work, such as a phone call. WFM uses a proprietary algorithm designed to distribute the backlog of deferrable interactions across the day in order to satisfy your service goal, which is expressed in minutes, hours, or days.
Spreading out the deferred work enables you to avoid spikes in workload forecasts when a contact center opens for the day, or during brief periods of high volume.
Multi-Skill Support
A multi-skilled contact center presents an opportunity for increased productivity.
An agent might be idle in a single-skill environment, because she cannot answer calls that are queuing for a particular activity/skill that she possesses—because the schedule prevents her from using that skill.
In a multi-skilled environment, the agent can use her additional skills to answer calls. A multi-skilled agent is qualified to work on multiple activities, and therefore can perform different types of work during a shift.
In a multi-skill environment, an agent can be available for multiple activities during any timestep. The agent can be scheduled to work on an activity for only part of a timestep, and only the fraction of the time period during which she works is counted.
Because of this, the value for staffing can be expressed as a fraction. For details, see Multi Forecasting Primer.