Purpose

Businesses need to make quick decisions and adjustments to fulfill the growing online demand. Previous studies examined various factors affecting the online sales performance of products such as books, electronics and movies; however, they paid limited attention toward the local brand clothing products. The current study investigates the importance of different kinds of seller-generated and consumer-generated signals such as price, discount, product ratings, review volume, review sentiment, number of questions and interaction between some of these factors for predicting the sales performance of clothing products.

Design/methodology/approach

The multiple linear regressions has been employed to investigate the influence of various predictor variables on sales performance. The study also examines the importance of these predictor variables by using different machine learning models, including random forest (RF), neural networks and support vector regression (SVR).

Findings

The findings of the study emphasize the importance of price and discount rates offered on the product. The quantitative characteristics of reviews, such as review volume and average rating, have been found to be more important predictors than sentiment strengths. However, the sentiment strength of reviews with higher helpfulness scores plays a significant role in predicting sales performance.

Originality/value

The study highlights the varying importance of seller-based and consumer-based signals in predicting sales performance. It also investigates the interaction effect of these two kinds of signals. The consumer-generated signals have been further divided into two components based on social influence theory, and the interaction effects of these components have also been examined.

中文翻译：

---

线上信号对本土品牌服装产品销售业绩的差异化影响

目的

企业需要快速做出决策和调整以满足不断增长的在线需求。以前的研究调查了影响书籍、电子产品和电影等产品在线销售业绩的各种因素；然而，他们对本土品牌服装产品的关注有限。目前的研究调查了不同类型的卖家生成和消费者生成的信号的重要性，例如价格、折扣、产品评级、评论量、评论情绪、问题数量以及其中一些因素之间的相互作用，以预测服装的销售业绩产品。

设计/方法/途径

多元线性回归已被用于研究各种预测变量对销售业绩的影响。该研究还通过使用不同的机器学习模型（包括随机森林 (RF)、神经网络和支持向量回归 (SVR)）来检验这些预测变量的重要性。

发现

该研究的结果强调了产品价格和折扣率的重要性。评论的数量特征，例如评论量和平均评分，已被发现比情绪强度更重要的预测因素。然而，具有较高帮助分数的评论的情绪强度在预测销售业绩方面起着重要作用。

原创性/价值

该研究强调了基于卖家和基于消费者的信号在预测销售业绩方面的不同重要性。它还研究了这两种信号的交互作用。基于社会影响理论，消费者产生的信号进一步分为两个部分，并且还研究了这些部分的交互作用。