Outbound call pacing method which statistically matches the number of calls dialed to the number of available operators

DETAILED DESCRIPTION An object of this invention is the provision of an improved call pacing method that can be used in a variety of calling environments.
Another object of this invention is the provision of a call pacing method that provides good agent utilization with small agent populations, and with low percentages of acceptable abandon calls.
Briefly this invention contemplates the provision of a call pacing method in which agent availability and call answers are based on probability statistics and the allowable maximum abandon rate.
A weight is assigned each agent engaged in a call or in after-call work.
This weight is a function of how long an agent has been engaged in a call or in after-call work, the length of adjustable time windows, and the maximum allowable abandon call rate.
The sum of the weights is a predicted number of engaged agents that will become free.
The length of the time windows is typically not the same for agents engaged in a call and agents engaged in after-call work and is not explicitly dependent on the interval between the dial of a call and when the call is expected to be answered.
Each time window is separately adjusted by the system; respectively lengthened when the abandon call ratio falls below the acceptable ratio and shortened when the ratio rises above the acceptable ratio.
Calls to be dialed are separated into classes based on the probability that a dialed call will be answered (e.
g.
previously dialed, busy responses will have a higher probability of answer than a previously dialed no answer response).
Separate data of answer vs.
dials are kept for each class.
Dialed calls that remain unanswered are further discounted as the unanswered status extends in time and the data indicates a diminished probability of being answered.
For each class, a number of calls to dial is calculated, based on the probability of an answer, in order to provide answered calls for the predicted number of idle agents without the predicted number of abandons exceeding the allowable abandon rate (as used herein, the phrase "predicted idle agents" (FS) includes engaged agents predicted to become available plus presently idle agents)