第三方在线购物平台中信誉对价格和销量的影响研究

发布时间：2018-04-16 18:55

本文选题：信誉体系 + 市场竞争　；参考：《哈尔滨工业大学》2016年博士论文

【摘要】：互联网的普及对用户的购物模式产生了巨大影响,越来越多的用户选择在网络购物。网上交易过程中,买家会面临更高的交易风险和更严重的信息不对称。为了减弱这些不利影响,众多第三方购物平台引入了信誉体系,从而增强买卖双方的信任、避免逆向选择。信誉已成为卖方传递质量信息的重要信号,成为买方识别高质量的卖家的重要依据。针对产品价格和销量这两个衡量网上卖家销售效果的重要因素,本文在以往研究的基础上,对信誉对价格和销量的作用进行深入探索。之前的研究利用不同网站的数据发现信誉对价格和销量影响模型的显著性并不一致,本文针对信誉对价格和销量影响有效性方面的不同结论,通过实证数据发现针对同一产品在不同网络平台中(淘宝网和e Bay网),信誉对价格和销量影响的显著性的确会存在差异。同时本文也指出该产品在两个网站市场竞争程度不同,继而构建了市场竞争对信誉与价格、销量关系的调节作用模型。本文指出即使在不同交易平台,信誉对价格和销量均有显著的影响,并指出竞争者数量会对信誉与价格、销量的关系起到调节作用。鉴于产品定价的动态性,本文收集了连续时间段内卖家的定价数据,探讨了信誉对卖家定价行为的影响。本文基于淘宝和天猫卖家对于刚上市产品的定价和价格调整行为,对比了不同信誉等级的卖家在新产品定价策略和价格调整策略方面的异同,本文发现低信誉卖家刚进入市场时更易采用渗透策略,其销售过程中的价格调整行为会受到高信誉卖家的影响。以往研究通常将销量视为连续型变量,本文考虑到销量的计数性特征,分别采用泊松回归、负二项回归和零膨胀负二项回归这三类计数型数据分析模型,对信誉与销量的关系进行拟合。本文发现产品销量具有过度离散、过度零的特点,零膨胀负二项回归模型能更好的拟合这类数据。数据结果显示信誉不但影响卖家销量的多少,也直接卖家销量成为结构零的可能性。此外,本文也发现不同地理区域的卖家销量会受所处区域的影响,但与信誉的影响作用不同的是,地理位置信息并不会影响卖家销量成为结构零的可能性。针对于动态的销量数据,本文发现传统的混合截面模型不能有效进行预测。为了动态地描绘信誉与销量的关系,也为了卖家在实际销售过程中更好的预测产品销量,本文通过非平衡面板数据,逐步构建了信誉和销量的固定效应模型。随后又利用零膨胀负二项回归模型,加入虚变量探讨信誉对销量的影响。本文发现信誉与销量的动态关系仍显著,同时信誉仍会影响销量结构零部分的预测,但预测的显著性有所下降。本文通过编写数据抓取程序,在淘宝网和e Bay网上获取了不同产品的销售信息,以及销售该产品的所有卖家信息。通过SPSS和SAS软件进行分析,逐步拓展了信誉对价格和销量的影响作用方面的研究。本文通过新颖的视角和方法,探讨了信誉与价格、销量动态和静态的关系。本文的研究成果在理论上,丰富了信誉和价格、信誉和销量的研究模型和理论;在实践中,对卖家定价决策的制定、产品销量的预测均有现实的指导意义。
[Abstract]:The popularity of the Internet has a great impact on the user's shopping mode, more and more users in online shopping. Online transaction process, buyers will face higher transaction risk and more serious information asymmetry. In order to reduce these adverse effects, many third party shopping platform introduces the credit system, so as to enhance the trust of buyers and sellers to avoid adverse selection, credit has become an important signal. The delivery of quality information, has become an important basis for the seller buyer identification for high quality products for the price and sales volume of the two measure online sellers sales effect of the important factors, based on the previous studies, in-depth exploration of the effects of reputation on prices and sales. Previous studies using different web site data found reputation on prices and sales significantly impact model is not consistent, the credibility of price and sales Conclusion the different influence validity, found that for the same product in different network platform through the empirical data (taobao.com and E Bay network), significant impact on the credibility of the prices and sales will indeed be different. At the same time this paper also points out that the product competition degree in two different market sites, and then set up the market competition of reputation and price regulation model of sales relationship. This paper points out that even in different trading platform, the credibility of the price and sales have significant effects, and pointed out that the number of competitors to reputation and price, sales relationship play a regulatory role. In view of the dynamic pricing, this paper collected the data of continuous time pricing in the seller, to investigate the effect of reputation on pricing behavior. In this paper, Taobao and Tmall for the Seller pricing and price adjustment behavior just listed products based on the comparison of different reputation The similarities and differences between the rating of the seller in new product pricing and price adjustment strategy, the penetration strategy more easily found low reputation sellers just entering the market, the price adjustment behavior of the sales process will be affected by the high credibility of the seller. The previous research usually sales as continuous variables, considering the characteristics of counting sales, using Poisson regression, negative two regression and zero inflated negative binomial regression two of these three types of count data analysis model, used to describe the reputation and sales. We found that the sales of products with discrete, excessive zero, zero inflated negative two regression model can better fit the data the results show that not only affect the reputation. The number of sellers sales, direct sellers sales also become zero possibility structure. In addition, this paper also found that different geographical areas and sales by the seller The regional influence, but influence and credibility of the effect is different, geographic location information and the seller will not affect the sales possibility to become zero. Sales data structure for dynamic, this paper found that the mixed section of the traditional model can not effectively predict. In order to describe the relationship between dynamic reputation and sales, but also to better in practice sales in the process of product sales forecast, the unbalanced panel data, and gradually build a fixed effect model of reputation and sales. Then using zero inflated negative binomial regression model two, adding dummy variables to investigate the effects of reputation on sales. We find that the dynamic relationship between reputation and sales are still significant, and will affect the reputation forecasting sales structure of zero part, but significantly decreased the prediction. In this paper, through the preparation of data capture program, Bay and E in taobao.com online access to different product sales Information, all sellers information and sales of the product. Through the analysis of SPSS and SAS software, and gradually expand the research effect of reputation on price and sales. This article through perspective and novel methods, discusses the relationship between reputation and price, sales of static and dynamic. The results of this study in theory, the rich the reputation and price, research model and theory of reputation and sales; in practice, the formulation of the pricing decision, product sales forecast has practical significance.

【学位授予单位】：哈尔滨工业大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：F724.6;F274

【相似文献】