通路管理機制研究-以Y公司為例

Managing Channel Profits – A Case Study of Company Y

楊莉珍(Li Chen Yang)

指導教授 ： 吳政鴻

國立臺灣大學/工學院/工業工程學研究所/碩士(2020年)

https://doi.org/10.6342/NTU202001388

摘要

企業為能提升市場佔有率與維持獲利，除了強化自身產品品質或技術外，還要能殷勤的服務客戶，並藉由客戶的意見反饋提升滿意度，增加銷售額。尤其近年來，在產品銷售活動中，不僅產業競爭日趨激烈，更因產業互聯網與區塊鏈的急速發展，導致市場供應鏈產生巨大的變化，企業經營時須不斷的面臨著直接與間接通路管理策略的選擇，如何在營運成本不斷上漲的微利時代下，突破重圍並提高整體市場佔有率，如何選擇高效益的行銷通路，將是影響企業獲利的重大因素，而該選擇何種適合通路及產品組合亦是企業目前所面臨的重要課題。
本研究運用多元線性迴歸分析與機器學習之XGBoost演算法協助探勘Y公司的歷史交易資料並建立預測模型，結果發現XGBoost建立預測模型的預測績效優於多元性線性迴歸的預測績效。另外運用XGBoost建立預測模型偵測出交易資料的離群值，由離群值比對原始資料並驗證歸納整理，追溯離群值形成原因，得知「產品組合的小類品項」與「客戶級價分類的不同」均會導致公司的可控淨利率轉變成負值，須對該銷售產品組合做調整與客戶分群加強管理，始能避免公司虧損，進而提升公司獲利。藉由本研究的離群值預測與分析，除可協助通路主管做有效管理，避免造成配合部門或業務人員的銷售績效偏頗，影響公司獲利外；亦可避免人工的解讀及判定減少管理上獲利的損失，而公司管理階層更可據此針對各事業部交叉比對個別客戶經營績效，做為未來規劃及調整公司整體通路策略的參考。

關鍵字

多元線性迴歸 ； XGBoost ； 預測模型 ； 離群值

並列摘要

In order to increase the occupancy rate for marketing and maintained its profit, companies should not only strengthen their product quality or technology, but also be able to serve customers attentively, and improve satisfaction while increasing sales through customer feedback. Especially in recent years, not only the keen industry competition but also the rapid development of the internet industry and blockchain had drastically changed in the market supply chain. The direct or indirect channel management strategies are what business enterprises will face these days. So how to break through the encirclement and improve the occupancy rate for marketing in the era of low profit will be the most important topic. Therefore, how to enhance the occupancy rate for marketing and choose a high-efficiency marketing channel will be the major decision that affects the profitability of the company and also be an important issue which channel and product combination should be selected for enterprises.
In this study, a multiple linear regression analysis and XGBoost algorithm were both used to explore the historical transaction data of Y company and established a prediction model. The results showed that the prediction performance of XGBoost model is better than that of multiple linear regression. In addition, XGBoost was used to establish a prediction model to detect outliers of transaction data. The outliers were compared with the original data and explained the causes of outliers. " Categories of product “and “Different Customer-level " caused the company's controllable net profit to negative. We can avoid the company's losses and increase the company's profit by adjusting the categories of product and customer groups. The results of outlier analysis in this study can not only help channel supervisors to do effective management, but also avoiding to affect the company's profit. Based on this result, the management of the company can compare the operation performance of individual customers of each business unit, and as a reference for plan and adjustment of the company's channel strategy in the future.

並列關鍵字

Multiple Linear Regression ； XGBoost ； Prediction Model ； Outliers

參考文獻

參考文獻

1.Capelleveen Guidovan, Poel Mannes, Mueller Roland M., Thornton Dallas, Hillegersberg Josvan(2016) ,Outlier detection in healthcare fraud: A case study in the Medicaid dental domain, International Journal of Accounting Information Systems,Vol. 21, Pages 18-31.

2.Cavusgil, S. T., Zou, S. (1994), Marketing strategy-performance relationship: an investigation of the empirical link in export market ventures. The Journal of Marketing, Pages 1-21.

3.Chatthipmongkol M., Jangphanish K. (2016),Factors Influencing Consumer Decision-Making Process of Thai Frozen Food Products, International Business management…, Volume 10, ,Pages 166-175.

4.Dong J.Q., Yang C.H.(2020) ,Business value of big data analytics: A systems-theoretic approach and empirical test, Information Management,Vol. 57, Issue 1, Pages 103-124.