消費者品牌移轉行為與市場佔有率之預測

Incorporating Market Share Forecasting with Consumer Brand Switching Behavior

本研究首先以 Logit 模式預測消費者之品牌移轉機率，再利用 Markov 模型進行未來市場佔有率之預測。藉由本研究模型，可以瞭解顧客品牌移轉的行為模式，並且預測公司未來之市場佔有率。本研究以遠傳電信公司資料庫資料進行實證分析，其主要結果為：第一，接受促銷活動的次數越多，其行動電話門號存活期限越長。因此行銷經理可致力於提高顧客參加促銷活動的機會。第二，以雙因子變異數分析結果顯示，遠傳電信應針對接受促銷次數較高的都會區族群，進行品牌忠誠度計畫；並且針對接受促銷次數較低的非都會區族群，加強促銷活動，以使其變成高忠誠度顧客。第三，在市場佔有率預測方面，利用 Markov-Logit 模型，對於預測一期、二期、三期之市場佔有率具有相當準確之預測能力，其平均 Theil’s U 分別僅為 0.0453、0.0751、以及0.0880，因此本研究所發展之模式具有一定應用價值。 The theoretical model developed by this paper is based on logit model for consumer brand switching probability and on the Markov model for market share forecasting. For policy-making point of view, the precise modeling of the consumer brand switching probability and market share forecasting. Moreover, the panel data (vs. questionnaire) used by empirical study of this paper could further capture the “true” consumption behavior. The results of this empirical study show that: first, the more promotion does the consumer participate, the longer the duration he is. Therefore, the marketing manager should focus on increasing customer’s participation probability. Second, consumer “location” and “consumer promotion participation” play intervening roles in duration time. Therefore, the manager should make efforts on the high participation -urban customers and lower participation -rural customers to effectively increase their duration time. Third, the Markov-Logit model developed by the paper could preciously forecast the next three periods market share resulting in are 0.0453, 0.0751, and 0.0880, respectively.