Scholarlink

<p>The sales performance of live streaming e-commerce streamers varies considerably, and understanding consumers’ shopping perceptions is critical for both platforms and streamers. We collected 600 live streaming e-commerce sessions and employed the multiple linear regression to examine the drivers of streamers’ sales performance from the perspective of consumer perceived value. In addition, as a supplementary validation, an eXtreme Gradient Boosting (XGBoost) model was applied to predict sales performance, and SHapley Additive exPlanations (SHAP) were used to interpret the corresponding influencing mechanisms. The results indicated that visual attractiveness, immersion, danmaku density, danmaku valence, perceived price, viewer identification, and streamer reputation have significant positive effects on sales performance. Conversely, access convenience and interaction rate show no significant impact. Furthermore, categorizing streamers into low-tier, mid-tier, and top-tier groups based on follower volume revealed distinct patterns in sales performance. Low-tier and mid-tier streamers benefit significantly from social identification and emotional interaction, whereas top-tier streamers rely primarily on price competitiveness. This study provides insights for streamers, platform managers, and multichannel networks (MCNs) to optimize their marketing strategies.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Exploring the driving factors of live streamers’ sales performance using explainable machine learning

  • 期刊论文

  • 发表日期：

    2026年2月28日

查看全文
  • Danni Zhang,
  • Xinyu Tao,
  • Caiming Huang

摘要

The sales performance of live streaming e-commerce streamers varies considerably, and understanding consumers’ shopping perceptions is critical for both platforms and streamers. We collected 600 live streaming e-commerce sessions and employed the multiple linear regression to examine the drivers of streamers’ sales performance from the perspective of consumer perceived value. In addition, as a supplementary validation, an eXtreme Gradient Boosting (XGBoost) model was applied to predict sales performance, and SHapley Additive exPlanations (SHAP) were used to interpret the corresponding influencing mechanisms. The results indicated that visual attractiveness, immersion, danmaku density, danmaku valence, perceived price, viewer identification, and streamer reputation have significant positive effects on sales performance. Conversely, access convenience and interaction rate show no significant impact. Furthermore, categorizing streamers into low-tier, mid-tier, and top-tier groups based on follower volume revealed distinct patterns in sales performance. Low-tier and mid-tier streamers benefit significantly from social identification and emotional interaction, whereas top-tier streamers rely primarily on price competitiveness. This study provides insights for streamers, platform managers, and multichannel networks (MCNs) to optimize their marketing strategies.