標題: 重複購買之銷售預測─以台灣線上光碟銷售網站為例
Repeat Sales Forecasting for CD online Shopping in Taiwan
作者: 吳全益
Chuan-Yi Wu
唐瓔璋
Edwin Tang
管理學院經營管理學程
關鍵字: 顧客關係管理;資料庫行銷;BG/NBD;RFM;重複購買;預測;customer relationship management;Database marketing;BG/NBD;RFM;Repeat Purchase;Forecast
公開日期: 2007
摘要: 近年來由於電子商務蓬勃發展,尤其是企業對消費者(B2C)市場更是競爭激烈,企業為了獲取更多的利益,無不想要擁有更多的客戶。然而因為網路世界的便利、隨時及無地域性的特性,使得消費者在取得資訊及進行交易的成本降低,消費者對於企業的忠誠度也大幅下降,電子商務企業的客戶流失率升高,客戶的消費行為越來越難以掌握及預測。然而企業的資源有限,無法對所有客戶投入行銷資源,因此顧客關係管理(Customer Relationship management)成為企業管理的主要課題,而其中最重要的關鍵如何能夠預測這些客戶的消費行為。 近年來行銷研究領域中一個熱門的議題就是如何利用已知的顧客購買行為(如一定期間內,顧客的購買次數、購買金額、最近一次購買日)等資訊,應用科學方法來掌握顧客購買行為的變化,運用不同的機率模型來直接算出每個客戶的活躍度及交易次數,而且這些模型是動態的,可以隨著時間及客戶消費行為變化直接反映在預測指標上。本研究採用Bruice 與Hardie兩位學者針對美國線上光碟片銷售網站CDNOW提出的BG/NBD模型,來預測台灣地區某線上光碟銷售網站消費者的重複購買行為,應用Microsoft的試算表軟體Excel來建立完整銷售預測模型,使用該網站的實際交易資料來驗證該模型對於銷售資料的適用性及預測能力,運用驗證資料來檢驗預測的準確性,並與一般常用的線性迴歸預測模型加以比較兩者在預測能力上的優劣。 本研究結果證明BG/NBD模型能夠很容易的在Excel上建立預測模型,並得到準確的銷售預測結果。本研究希望藉由該模型在整體銷售預測上的準確性,能夠協助企業預測客戶重複購買行為及計算顧客未來價值。
Along with the development of marketing, the companies notify the fact that the customer is important to the market competition and their future development. The customer relationship management has become a very important aspect of business management. The core of customer relationship management is customer behavior prediction. The major objective of this research is to assist companies in forecasting the repeat purchase behavior of customer. Aiming at the BG/NBD prediction model (Fader and Hardie 2001), this research applying a simple stochastic model in conjunction with the online CD retailer in Taiwan .The data on the purchasing of CDs at the online retailer X-Company is used as empirical analysis. This research tries to provide modeling exercise to forecast medium-term aggregate CD purchasing by a cohort of new customers. We modeled monthly sales using finite mixture of beta-geometric distribution with a time-varying function to capture the nonstationarity in repeat purchase. In the result this model can easily be implemented within spreadsheet software like Microsoft Excel .The parameter calculation is very efficient and easy by using “Solver” function in Excel. This model provides the ability to describe the underlying sales patterns and produces an excellent purchasing forecast for the X-Company.
URI: http://140.113.39.130/cdrfb3/record/nctu/#GT009474517
http://hdl.handle.net/11536/82652
Appears in Collections:Thesis


Files in This Item:

  1. 451701.pdf

If it is a zip file, please download the file and unzip it, then open index.html in a browser to view the full text content.