Predictive Outbound Algorithm And The Call System Prototyping

As market competition intensifies, enhancing the electronic sales and service capacity has become a problem enterprises to achieve strategic transformation must face. More and more companies started to build their own automatic outbound system to provide proactive services to customers. Through a standardized and effective telemarketing, telephone surveys and customer visits and other services, companies can promote new business, carry out comprehensive and effective collection of customer data, deliver products, services and promotions to customers accurately and timely and facilitate customer orders. Ultimately achieve the goal of enhancing customer service level and customer value creation.The core of automatic outbound system is predictive outbound algo rithm, predictive outbound algorithm has to increase agent utilization, and also decrease call loss rate to avoid harassment which influence customers' satisfaction and loyalty. So how to balance agent utilization and call lose rate becomes an core and the most important performance indicators to evaluate predictive outbound algorithm. In this paper, the main line in predictive outbound algorithm based on dynamic statistics, the main results include the following three aspects:First, it's can concluded that the service time approximately obeys negative exponential distribution, the ring time approximately obeys the generalized Poisson distribution and the biggest factor of the hit rate of phone with the mean of the other parameters is the all times of every day and so on through the statistical analysis of the actual manual outbound dialing data. According to these conclusions, people can build a more reasonable mathematical model and choose a simple and efficient statistical method to count the historical data. It provides a solid theoretical foundation and guidance to the establishment of predictive outbound algorithms.Then, aimed at the shortcomings of current predictive outbound algorithm, we propose an algorithm that divides predictive outbound algorithm into two parts, predictive outbound algorithm and dynamic statistic. We provide optimal methods for both parts, avoiding the bottleneck caused by only optimizing one of them. Dynamic statistics allows to employ various kinds of strategies according to the current state and historical data without concern of how the algorithm are implemented or whether it's useful, it can also generate random numbers in accordance with common distributions, like exponential distribution, Poisson distribution, normal distribution, C hi-square distribution and gamma distribution etc., based on the theoretical model built by the algorithm. Predictive outbound algorithm module provides an effective algorithm based on queuing theory model and so on, increasing agent utilization and decreasing call lose rate at the same time.Finally, we set up a call center outbound prototyping system which can simulate call center automated outbound process, for testing the effectiveness of various types of predictive outbound algorithm. The system can simulate call, acceptance, services, and hang up the process, while dynamically count various parameters and real-time monitor agent utilization and call lose rate. In addition, the system can not only simulate the actual data, but also to generate simulated data in order to simulation comparison, thus having great significance for companies, testers and researchers.