行動電話業客服話務預測之研究-以某客服公司為例

A Study on Forecasting Incoming Calls to the Customer Service

張世青(Shih_Ching Chang)

指導教授 ： 古思明 博士 ； 龐金宗 博士

元智大學/資訊學院/資訊管理學系/碩士(2003年)

摘要

近年來，由於企業對客戶服務品質的重視，扮演與客戶接觸介面的客服中心相形日趨重要。本研究對象為一專業客服公司，其主要業務為國內行動電話電信業者的客戶服務業務，由於行動電話用戶在近幾年內快速增加，伴隨著面臨話務量快速成長的情況，公司的組織也隨著服務人員增加而不斷的擴充，至今將近有兩千名員工。管理者必須在有限的預算內，做最妥善的資源分配包括：需要多少專員執機、招募多少人員、建置多少執機的席位數、電話線路需租用數量等，最終目標就是要提升客戶滿意度，而要能精確地預測未來的話務量是邁向成功的第一步。
本研究目的要建立預測模式，預測每日TCC行動電話業客戶服務電話的來電話務量，以季節性Box-Jenkins時間序列方法建立預測模式，話務歷史資料採用期間：從民國八十九年一月至九十年六月止，以百分之九十五的機率區間預測未來一年，亦即從九十年七月到九十一年六月底的話務量，並與實際觀察值比較，且就模式作分析討論。
為了增加預測的精確度，從訪談得一假設在過去1、2個月新成長的客戶數將對當月的話務量變化造成影響，所以從交通部統計處蒐集TCC客戶成長月資料，資料區間從民國八十九年五月至九十年十二月，欲建立電信公司的用戶數成長數與服務電話的話務量變化之轉換函數模式，但是以交叉相關函數方法顯示的結果卻是未來三個月預期的客戶成長數將會對當月份的話務量變化造成顯著的影響，以此作為本研究的發現。

關鍵字

時間序列 ； 話務 ； 預測 ； 客服中心 ； 轉換函數模式

並列摘要

In the recent years, increasing numbers of the enterprises are investing in customer service implementation strategies and practices. Hence the call center plays a important role in an interactive channel between the enterprises and their customers. In this research, we use case study to the company supply teleservices and technologies, the major account of this company is TCC telecommunication cooperation. The primary purpose of the call center could be to receive calls for handling customer service questions and issues that typically have been placed to an 080 number or 888 short call. Followed increasing drastically numbers of the customer using the cellular phone, the traffic of incoming call to the center also increased and the organization had faced to extend. Until now there are almost 2000 employee work for the center. The managers have to distribute properly resource in the budget, included in sufficient personnel to answer calls, sufficient network service lines or trunks, and arrangement of center terminal equipment. The goal is to increase customer satisfaction and the first step to reach the goal is forecasting traffic accurately .
In this paper, forecasting model is constructed for incoming call volume to the center . Using an iterative procedure developed by seasonal Box-Jenkins approach. Data covering the period of January 2000 through June 2001 were used to develop the model. Forecasting with 95 percent probability limits were calculated for one year from July 2001, and were compared with the actual observations, the properties of the model are discussed in detail.
In order to improve the accuracy, an idea raised on the interview is how the monthly increasing numbers of the customer impact the traffic of incoming call to the center. Collecting the monthly data from the Ministry of Communication, data covering the period of May 2000 through December 2001. Originally constructed the forecasting model with transfer function, the cross correlation function technique is demonstrated to the customer-traffic relationship, with particular focus on the customer lag structure. The result shown that the numbers of the customer in the third month will significantly effect upon the present call volume .

並列關鍵字

Time series ； Traffic ； Forecasting ； Call Center ； Transfer Function.

參考文獻

5.林琬儒，「電話服務中心之服務品質分析：以中華電信障礙服務為

6.宋彥青，「行動電話消費行為研究－業者、費率方案、與使用量混合

月，pp.102-107。

11.Andrews, Bruce H. and Cunningham, Shawn M. 1995, ”L.L.

Bean improves call-center forecasting,” Interfaces, Vol.