数据分析技术驱动短视频营销的影响因素分析——基于多元线性回归分析模型

Analysis on Influencing Factors of Short Video Marketing Driven by Data Analysis Technology —Based on Multiple Linear Regression Model

摘要: 为挖掘影响短视频营销效果的关键因素，量化各因素影响程度并为企业优化营销策略提供实证参考，本研究以抖音、快手、微信视频号三大平台快消、美妆等四大行业的420个企业营销短视频为有效样本，选取内容质量、发布策略等6个可量化的核心自变量，构建融合曝光、互动、转化、复购引导的营销效果综合值作为因变量。通过Z-score对所有自变量进行标准化数据预处理后，分别构建多元线性基础回归模型与加入成本投入二次项的修正回归模型，经拟合度、显著性及多重共线性检验确定最优模型。研究结果表明，内容质量、平台算法适配等6个因素均与短视频营销效果呈显著正相关，影响程度从大到小依次为内容质量、平台算法适配、发布策略、互动运营、成本投入(线性项)、达人合作；成本投入存在边际效应递减，标准化阈值为0.8539。本研究构建的回归模型拟合效果良好，可用于短视频营销效果预测与策略优化，为企业开展短视频营销提供实操性指导。

Abstract: To identify the key influencing factors of short-video marketing effectiveness, quantify their impact and provide empirical references for enterprises to optimize marketing strategies, this study takes 420 enterprise marketing short videos as valid samples from four major industries including FMCG and cosmetics across three platforms: Douyin, Kuaishou and WeChat Channels. Six quantifiable core independent variables such as content quality and publishing strategy are selected, and a comprehensive marketing effectiveness index integrating exposure, interaction, conversion, and repurchase guidance is constructed as the dependent variable. After standardizing all independent variables using Z-score for data preprocessing, a basic multiple linear regression model and a modified regression model with the quadratic term of cost input are established respectively. The optimal model is determined through goodness-of-fit, significance, and multicollinearity tests. The results show that all six factors, including content quality, platform algorithm adaptation and etc., have a significantly positive correlation with shortvideo marketing effectiveness. The order of impact from greatest to least is as follows: content quality, platform algorithm adaptation, publishing strategy, interactive operation, linear term of cost input and influencer cooperation. Cost input exhibits diminishing marginal effects, with a standardized threshold of 0.8539. The regression model established in this study achieves satisfactory goodness of fit and can be applied to predict shortvideo marketing effectiveness and optimize strategies, providing practical guidance for enterprises in short-video marketing.