基于单品SKU样本的跨境电商销量模型分析  

Analysis of cross-border e-commerce sales prediction model based on single product SKU sample

作　　者：黄晓卫 杭天 汪霞 Huang Xiaowei;Hang Tian;Wang Xia(Jiangsu SOHO International Group Co.,Ltd.,Nanjing,Jiangsu,210012;Holly Futures Co.,Ltd.,Nanjing,Jiangsu,210004;School of Economics,Wuhan Business University,Wuhan,Hubei,430056)

机构地区：[1]江苏苏豪国际集团股份有限公司,江苏南京210012 [2]弘业期货股份有限公司,江苏南京210004 [3]武汉商学院经济学院,湖北武汉430056

出　　处：《市场周刊》2025年第7期83-87,共5页Market Weekly

摘　　要：电商企业的单品销量通常具有数值小、季节性周期不强、易断货下架的特点,与电商大环境和产品所属行业的周期关系并不明显。为助力中小电商企业的数字化决策,文章运用SKU层级的销量数据,构建了Hedonic Price Theory(HPT)销量模型和Multivariate Poisson Autoregressive(PAR)销量模型。上述模型不仅可以实现销量的预测,还可在同一框架下定量估计广告预算、价格弹性、节日、断货等对销量的影响。除此之外,研究还涉及节日旺季的识别和广告的内生性问题研究,以期为电商的销售预测、销售分析提供参考。E-commerce firms'single product sales are typically characterized by low values,weak seasonal cycles,and frequent outof-stock and off-the-shelf situations,and in most circumstances,those SKU sales are not strongly tied to e-commerce sector statistics or overall economic indicators.This study uses sales data to fit the Hedonic Price Theory(HPT)and Multivariate Poisson Autoregressive(PAR)sales models,with the goal of assisting small and medium-sized e-commerce businesses in making decisions.The models above not only anticipate sales volume,but they also quantitatively qualify the effects of advertising budget,price elasticity,festivals,and out-of-stock on sales volume within the same framework.In addition,this study identifies peak holiday seasons and investigates the endogeneity problem of advertisement in order to provide a reference for e-commerce sales forecasting and analysis.

关 键 词：销售预测 销售分析 电商 SKU 

分 类 号：F713.3[经济管理—产业经济]