零售大數據挖掘與區隔:社群口碑與顧客決策分析

在高度虛實整合，消費者更主動掌握訊息，以及移動技術快速發展下，提供了嶄新的購物消費體驗，而讓零售通路經營的挑戰不斷提高，然而善用大數據瞭解消費者行為，並透過分析找出顧客的決策，並進行行銷預測。本研究利用質性研究與量化研究，質性研究透過美妝零售社群網站Fashion Guide之消費者評論，進行評論蒐集，計算評論值與情緒值，運用內容分析，透過消費者使用經驗發現影響消費者情緒與評論之相關字詞。量化研究，首先，問卷設計與問卷發放，所採用的分析方法包括信度分析、因素分析、集群分析、決策樹分析。
本研究探討零售社群口碑與顧客決策分析，研究結果得知質性研究，影響消費者情緒之關鍵字詞為「喜歡」產品、「擔心」產品適合性等情緒詞，評論之關鍵字詞為使用效果「清爽」、產品品質「不錯」等評論字詞。量化研究，透過市場區隔決策樹分析，購買動機目標變數發現女性消費者，年齡二十六至三十五族群，在選購保養品時會依據核心功能而進行選購，消費者每年平均購買次數四到五次，購買動機為風格象徵導向，針對此族群的女性消費者進行風格象徵廣告投放。購買次數目標變數，消費者每年平均購買次數八次以上，花費金額為四千以上，針對此消費族群可以在先在百貨公司、藥妝店、直銷方式、大型量販店等門口進行廣告文宣傳單發放，針對此族群進行相關商品廣告投放在社群網站針對風格象徵或享受樂趣的廣告。

With a high degree of virtual reality integration, consumers are more active in mastering information, and with the rapid development of mobile technology, it provides a brand-new shopping consumption experience, which makes the retail channel operation more challenging. However, we should make good use of big data to understand consumer behavior, find out customers' decisions through analysis, and make marketing forecasts. In this study, qualitative research and quantitative research are used. Through consumer reviews of fashion guide, a beauty retail community website, reviews are collected, comments and emotions are calculated. Content analysis is used to find out the relevant words that affect consumer sentiment and evaluation through consumer experience. Quantitative research, first, questionnaire design and questionnaire distribution, the analysis methods used include reliability analysis, factor analysis, cluster analysis, decision tree analysis.
This study explores the relationship between word-of-mouth and customer decision-making in retail communities. The results show that the key words influencing consumers' emotions are "like" products and "worried" about the suitability of products, and the key words of reviews are "refreshing" and "good quality". Through the analysis of market segmentation decision tree, we find that female consumers, aged from 26 to 35, choose maintenance products according to their core functions. The average number of purchases per year is four to five times. The purchase motivation is style symbol oriented, and the female consumers of this group carry out style symbol advertising. The target variable of purchase times: the average number of purchases per year is more than eight times, and the cost is more than 4000. For this group of consumers, advertising can be carried out at the door of department stores, cosmetology stores, direct sales methods, large-scale mass stores, etc. for this group of consumers, advertising related products should be put on the community website aiming at style symbols or enjoying fun.