- 學位論文
社群使用者黏著度遷徙路徑分析─以Facebook為例
The Analysis of User Stickiness Migration Path – An Empirical Study of Facebook
摘要
網路時代來臨，造就了許多社群媒體的興起，提供網路使用者之間對話互動，透過互動來產生內容，社群媒體的出現大大地改變現代人的生活型態和消費者行為，社群媒體上出現的網路口碑訊息則成為了重要的購買決策參考來源，因此越來越多的企業透過此管道來與消費者進行溝通，推廣品牌及拓展商務。 對於社群媒體經營而言， 如何留住社群用戶才能吸引更多企業投入廣告，提高黏著度(Stickiness)便是首要課題；第二，為了解社群用戶，針對特定目標客群進行溝通，讓廣告投放更加精準，協助企業配適最佳的廣告，使得廣告的投放能夠獲得最大的效益。 本研究利用ComScore資料庫2012年Facebook瀏覽紀錄，以消費者網路造訪行為－網站停留時間(Duration Times at Site)以及造訪次數(Visit Time)定義出使用者黏著度，再以馬可夫鏈模型為基礎對此進行路徑遷徙分析，描繪其動態性。以層級貝氏Probit估計出個人的移轉機率矩陣，嘗試捕捉不同使用者的異質性。根據保留樣本進行擊中率比較，確立模型效度。並建立六條具有行銷意義的遷徙路徑，分析不同路徑的使用者特性，最後再以馬可夫鍊之狀態機率極限預測最終落點。本研究之結果首次成功地提供一個預測模式可以精確地估計使用者未來黏著狀態，並以此結果研擬出行銷策略。
並列摘要
Social media has greatly changed people’s life style as well as consumer behavior since the modern development of the Internet. Based on this fact, electronic word-of-mouth (EWOM) on websites plays a main role in consumers’purchasing decision as a reference resource. Therefore, a significant number of companies has started to promote their products, communicate with their customers, and develop other business service by social media. For social media, first, due to the fact that the revenue of social media comes mainly from advertising, how to maintain the stable and attract more number of community users has become one of the most important issue. Second, in order to focus on the specific Internet community users, how to enhance the effectiveness of advertising raises another critical topic. For the first part of the research, we use ComScore database to define Facebook user’s stickiness according to user’s website browsing behavior and apply Markov Chain and Hierarchical Bayesian Model to build individual user’s Transition Probability Matrix to predict user’s stickiness state. For the second part, 6 migration paths with highly significance of marketing are set up to analyze the relationship between transition paths and demographic variables. As a result of this research, we successfully demonstrate a useful model that can precisely estimate user stickiness state, which is able to further assess marketing strategies.