|
Forecast
The forecast predicts the number of incoming calls that the call center will receive during a specified time interval for the requested scheduling period. For example, we use 30 minutes as our basic time interval so that the week, our requested scheduling period, is split into 336 time intervals.
The forecast is based on past history, day of the week, time of year, and holiday data, if any. Certain parameters are considered when building the forecast including:
- How far into the future the user wants to schedule?
- What data is being filtered out?
- What is the chance of a person who hangs up calling back?
Conversion
This step takes the expected number of incoming calls and figures out the number of agents needed to properly service each of these calls. The output of this phase, the demand graph, is the number of agents needed to fill each basic time interval for the entire planning period. A mathematical formula from the Queuing Theory is used to compute the required number of agents (the demand graph) from the estimated number of incoming calls. The main inputs to this formula are:
- Service level to be achieved. This is measured in percentage of calls that are answered in less than T seconds, where T is a measure for reasonable response time.
- Distribution behavior of incoming calls.
- Average and variance of the duration of one call.
Building shift duties
| |
|
