影響直播串流平台觀眾購買汽車周邊商品關鍵因素

Abstract

隨著網路行銷日益普及，消費者使用串流平台，於直播中購物的金額大幅攀升。汽車周邊商品具多樣性，能夠滿足各種不同市場區隔中，車主的需求和偏好，並有效提昇汽車的功能性、外觀、安全性或便利性，於部分國家或經濟體串流平台之銷售額日增。雖然汽車周邊商品之商機龐大，且已經擴散至如中國大陸等新興市場之串流平台，但於大部份國家之市場仍然處於萌芽期，因此，行銷相關之議題非常值得關注與研究，唯目前少有學者著墨。因此本研究擬分析消費者透過直播，購買汽車周邊商品的行為模式，並探討影響消費者透過直播購買汽車周邊商品的關鍵因素。本研究首先爬取與汽車周邊直播銷售相關網頁，並以基於隱含狄利克雷分布 (Latent Dirichlet Allocation，LDA) 之主題模型分析（Topic Modeling），探勘網頁中蘊含之主題；其次，以階層式集群分析法（Hierarchical Cluster Analysis）分群主題後，將各群主題歸入技術接受模式的構面，並以偏最小平方結構方程模型（Partial Least Squares Structural Equation Modeling，PLS-SEM）驗證各路徑顯著與否。除此之外，本研究亦同時邀集專家，以基於決策實驗室分析法（Decision Making Trial and Evaluation Laboratory，DEMATEL）之分析網路流程（DEMATEL based Analytic Network Process，DANP），推衍影響直播串流平台觀眾購買汽車周邊商品之關鍵因素，並比較社群網路探勘與專家意見法兩者結果之差異。依據實證研究，與購買意願相關係數最高的變數為「推廣效果」，分析網路爬取資料，與專家提供之意見一致，共同認為「推廣效果」為影響消費者透過直播，購買汽車周邊商品之最關鍵要素。本研究之結果，可作為未來制定直播行銷策略之基礎，發展之方法，亦可作為訂定發展其他新興網路媒體策略之用。

As online marketing continues to gain prominence, there has been a marked escalation in the expenditure of consumers on purchases made via live streaming platforms. Automotive peripherals and accessories boast a wide variety, capable of satisfying the diverse needs and preferences of car owners across different market segments. They effectively enhance a vehicle's functionality, aesthetics, safety, and convenience. In some countries or economies, sales figures on streaming platforms for these products are growing daily. While the business opportunity for automotive peripherals and accessories is vast and has expanded to streaming platforms in emerging markets, such as Mainland China, the market in most countries remains in its infancy. Thus, marketing-related topics in this area are very much worth attention and research. However, there's been limitedscholarly attention to this subject. This study aims to analyze the behavioral patterns of consumers purchasing automotive peripherals and accessories through live streaming and to delve into the key factors influencing their buying decisions on these platforms.This study initially crawled web pages related to live streaming sales of automobile peripherals and accessories and employs Topic Modeling based on Latent Dirichlet Allocation (LDA) to mine the underlying themes from these posts. Subsequently, the Hierarchical Cluster Analysis method is used to categorize these themes under the dimensions of the Technology Acceptance Model. The significance of each pathway is then validated using the Partial Least Squares (PLS) Structural Equation Modeling (SEM), or PLS-SEM. In addition, experts are invited to apply the Decision Making Trial and Evaluation Laboratory based Analytic Network Process (DEMATEL-based ANP) to deduce the critical factors affecting the audience's purchase of car peripherals on live streaming platforms. The study further compares the results from social network mining and expert opinion. Based on empirical research, the variable with the highest correlation to purchase intention is "promotional effectiveness." Analysis of data scraped from the internet, consistent with expert opinions, unanimously points to "promotional effectiveness" as the key determinant influencing consumers to purchase automotive-related products through live streaming. The findings of this study can serve as a foundation for formulating future live streaming marketing strategies. The methodology developed can also be applied to the strategy formulation for other emerging online media platforms.