最適促銷組合之策略 – 建構於馬可夫轉換模型

Abstract

本論文主旨在於研究，如何在競爭激烈的快速流通產品產業中建立促銷組合之最佳策略，在文中，我們利用品類管理的觀點分別檢視不同的品類、包裝以及通路，並且使用馬可夫轉換模型將銷售分為促銷期間以及非促銷期間衡量以衡量跨期促銷投入的效果。首先我們利用模型決定衡量過去促銷投入的標準，進而以對過去促銷投入的評鑑作為未來預算投入的準則，解決品牌經理或產品經理過去以直覺、過去經驗、或隨機決定預算分配的困境；我們利用馬可夫轉換模型所建立的假設，尋求未來進行促銷投入時，促銷期間與非促銷期間的比例，以及其間隔時間的長短、促銷次數等，；最後我們利模型所顯現的結果歸納出未來廠商進行策略決策時的準則。 我們使用目前世界最大的食品公司過去39個月的銷售以及促銷投入的資料，進行行銷組合之最佳策略之研究。數據結果顯示，過去的促銷投入的次數過多，且並未在刺激銷售上有明顯的效果；此外我們發現在擬定促銷方案時，應當考量產品銷售量的多寡，使用不同的促銷型式。最終的目的在於提供品牌經理以及產品經理一個具理論基礎的解決方案。

This paper investigates the optimal sales promotional strategy for the fiercely competitive FMCG (Fast Moving Consuming Goods) industry. We propose a Markov Switching Autoregressive model that incorporates AR(1) retailing demand process to capture nonlinear structure among promotional budget allocation, evaluation of promotion performance, and optimal promotion frequency within a given time span. The past promotion investment is evaluated first by comparing the changes in promotional budget allocation. We then apply Markov switching feedback rules to figure out the proper length of equilibrium state with and/or without promotion. Finally, effective decision rules on magnitude, duration, and frequency of promotional strategy are induced. We apply three product categories with 39 months time-series data from a multinational packaged food company. The result shows that most past decisions on promotional budget allocation are non-optimal – most promotion investments were either extended too long or allocated too low in stimulating sales. Implications for the brand- or category- manager in removing those non-optimal promotional policies are suggested.