The Prediction Research Of Customer Service Call Volume For HP China Desktop Business

Call center business was started in civil airline industry about 40 years ago, and it was introduced into China in 1990 s. With the rapid growth of Chinese economy, business competition and the improvement of customer service sense etc, call center business was developed very fast in the last ten years in China.There are two business models for call center, ‘self-built’ and ‘outsourcing’. But no matter in which model, they are facing two major challenges in call center operation management, volume imbalance and high percentage of labor cost. Especially in outsourcing call center, with more and more intense competition, how to balance the utilization of labor resources and the service level of customer service, and then better control the cost of operation especially labor cost became one of the major core competence. But in order to better control the labor cost, the precisely call volume prediction is a key precondition.This article introduces the call center queue theory at first and its widely implementation in call center operation management. Then it elaborates the business background and the current operation of HP China desktop after sales customer service project. As it’s an outsourcing project, the article analyzes the key factor of making profit. It comes out that under the same customer service level, how to reduce the labor cost, become the key factor to increase project margins. So the precisely call volume prediction becomes an import precondition.After analyzing the actual call volume and fluctuation pattern of 2009 and 2010, this article used the segmental regression model to predict normal days and used statistic model to predict holidays. And based on the call volume of 2009, it created the prediction model for 2010. It also verified the model self and its prediction capability.During the validation, it found that although the trend pattern of call volume is similar between different years but current approach will encounter large gap when the total volume becomes huge difference. So the article adjusted the current prediction model a little bit, and it calculated the fluctuate ratio of last three month, revised the previous call volume based on this ratio and then recalculated the regression model to predict the volume of next month. After verified this new approach, the result turned out to be more precisely.At last, the article proposed several operation suggestion to this project based on the prediction model.