應用決策樹探討研究所補教業者之電話行銷策略

Applying Decision Trees to Explore the Telemarketing Strategies of Cram Schools for the Graduate Institute

近年來台灣因受少子化與高學歷失業率居高不下，導致大學碩士班研究所報考人 數屢創新低，研究所補教業者現行作法，多以密集的電話行銷，來接觸大學在學生的 潛在顧客，並進行課程商品的銷售及服務解說。然而業者若沒有一套具系統性的方 法，來執行電話行銷策略，易造成效率與效能不彰，且對無升學意願的學生亦會造成 困擾，以致影響補教業者的形象。因此如何在眾多競爭者當中，如何有效地從大學在 學學生名單中挖掘出具有潛在升學意願的學生，實為業者很大的挑戰。本研究首先以 計畫行為理論為基礎，以結構式問卷調查高雄地區有/無升學意願之大四學生，了解 這些學生對於進修研究所之「行為態度」、「主觀規範」、「知覺行為控制」及「人口統 計變數」等的關鍵因素。研究結果發現，影響進修研究所意願的關鍵因素，包括「行 為態度」中的「完成學位能否增加收入」與「是否能藉由師長擴展人脈」；「主觀規範」 中的「家人與同儕是否認為該學生應繼續念研究所」；「知覺行為控制」中的「自己是 否有信心通過研究所考試」、「家中是否有足夠經濟條件支持該學生升學」與「是否有 信心面對未來研究所之課業」。其中以「是否有信心面對未來研究所之課業」為最關 鍵的影響因素。其次，應用決策樹所建構之升學意願預測模型，找出了10 條大學生 升學意願與否之規則型態，將規則型態再進一步區隔成5 個有升學意願，及5 個無升 學意願的族群，根據族群特徵加以描述及命名後，針對上述族群提出對應的電話行銷 策略與建議。研究結果可提供補教業者在執行市場區隔及制定電話行銷策略之參考。

Due to the lowering of the market needs and the surroundings of fierce competitions， the question of how to find out potential students with intentions for higher education has become the greatest challenge. To understand the factors of Attitude toward the Behavior， Subjective Norm， and Perceived Behavioral Control， this study employs the Theory of Planned Behavior as a framework and uses structured questionnaires in investigating both collegiate senior students with and without intensions for higher education. The results find the key influential factors include: ‘whether the degree can bring more salary’ and ‘whether the personal network can be expanded’ within the Attitude Toward the Behavior category; ‘whether family members and friends think it is worthy to go to the graduate school’ in the Subjective Norm; ‘whether I have the confidence to pass the test’， ‘whether there is enough financial support from the family’， and ‘whether I have faith for handling the school work’ within the Perceived Behavioral Control category; ‘whether the current school is a national one， private or vocational one or a normal one’ among the demographic statistical variables. Among these factors， the most important factor is ‘whether I have faith in handling the school work’. Secondly， this study uses the decision tree in constructing prediction model of intension for higher education， and ten types of the rules could be found. Types of the rules then are further grouped into 5 with the intension for higher degree and 5 without the intension for higher degree respectively. Based on the characteristics and the description of the groups， suggestions and marketing strategies were then proposed and given. Finally， the results of the study can play the role of being the reference for cram schools when conducting market segmentation and telemarketing.