竞价排名绩效及其影响因素的实证分析
|
摘要
从1970年,互联网应用开始至今,互联网技术经历了疯狂非理性的发展阶段,技术重整阶段,而进入到了今日仍在深刻演变并带来深远影响中的web2.0时代。在这40年中,我国互联网经济的发展势不可挡,2005年至2010年期间,国内互联网经济的发展速度远远超过国内GDP的增速。互联网应用对国内经济的发展产生了深远的影响。而其中的网络搜索技术,更是在近年内,跟随互联网应用的强劲势头得到了充分的发展。目前我国搜索引擎用户占全体互联网用户总数的比例早已超过了70%,用户数量也已突破2亿大关。目前即时通信服务,电子邮件服务和搜索引擎已经成为现今互联网的三大基础应用。借助网络搜索平台技术的迅速发展,竞价排名搜索这一新的网络推广模式从2000年开始应用至今,以其推广地高精准度,低进入门槛,和成本可控等优势,迅速的虏获了各类商家。目前竞价排名搜索成为国内外各大搜索引擎和行业门户网站的运营主要收益模式。
在此背景下,本文中以淘宝网直通车这一竞价排名搜索栏目的数据为基础,搜集了广告主真实的成本和收益数据进行实证研究,以弥补以往研究在此方面的不足,真实地估计竞价排名的实际效果。本论文通过分析消费者在使用竞价排名搜索应用时的不同行为流程,提出了商家直接绩效和溢出绩效两个新概念,研究在竞价搜索排名推广模式中,跨类溢出绩效是怎样产生的。并建立计量模型研究最初的竞价排名搜索直接绩效与溢出绩效之间的存在关系以及影响溢出绩效和直接绩效的影响因素。
本论文内容首先将在文献综述部分介绍目前国内外对于竞价排名搜索的相关理论研究,以及本论文所涉及领域相关的理论背景。然后通过数据背景部分说明此次研究所使用数据的来源和特点,详细说明数据中各个变量的具体含义,构成和特点以及对数据进行的描述性统计分析结果发现。在之后的假设与模型设计阶段将把这个课题分为三个部分：总点击量绩效模型、直接销售数量绩效模型和间接销售数量绩效模型,展示此次研究的计量模型设计,并在之后对实证数据的回归结果进行分析。最后,根据研究结果中的结论提出针对于广告主商家在关键词和竞价策略选择方面有指导意义的管理应用建议,并对此次研究作出总结、提出需要进一步改进的方向,且.为未来更深入的研究提出建议。
本文涉及的实证研究,使用来自淘宝网直通车栏目下的时间跨度达60天1105个商家共40162条实际销售面板数据进行研究。并主要得出以下几点结论：1.商品信息总点击量直接影响商家的直接销售绩效,而直接销售绩效会对间接溢出绩效产生正相关的作用。2.商家对关键词的选择会对商家各类绩效产生不同的影响：a.关键词的长度对总点击量绩效有负相关的影响效果,但会使得直接销售绩效正向增长。b.关键词中的品牌信息,在被关联商品为具有高品牌影响度性质的条件下,会对商品信息的点击量产生正相关的影响,而对于直接和溢出间接销售绩效则有负相关的作用。c.关键词中的促销信息会对商品信息的总点击量绩效产生显著正相关的影响。3.商家的信誉级别因素会对商家的直接、溢出间接销售绩效产生重大的影响。信誉级别越高,销售绩效越好。4.在该数据来源商品特征环境下,商品的价格因素对商家各类绩效影响均不显著。
综合以上的研究结果,本文针对商家参与竞价排名搜索推广中所涉及的三个选择策略提出了以下一些启示：1.增加与商品关联的竞价关键词的数量比提升单一关键词竞拍价格更能带来高的实际收益。2.选择竞价关键词时应当推广商品的特征属性以及推广目的,选择具有不同属性特征的关键词,获得不同的绩效影响。3.在选择推广商品时,应当选择能引致关联消费的商品,通过增加直接销售绩效,进而使得间接商品的销售绩效得到提升。
Since 1970, the start of Internet applications, Internet technology has undergone a crazy non-rational stage of development. After went through the re-engineering phase, it now has entered into the far-reaching and deep evolution phase of the web2.0 implications era. In the last 40 years, the economic development of the Internet in China was overwhelming. From 2005 to 2010, the development rate of the Internet economy was much faster than domestic GDP growth rate. Internet application development has made a profound impact on the domestic economy. Especially in recent years, the web search technology has been fully developed. Currently the user amount of search engines has taken more than 70% of the total number of all Internet users. The number of search engine user has exceeded 200 million. The current instant messaging service, e-mail services and search engines have become the three major applications of the internet. Since 2000, first launched, the sponsored search technology which was carried by the web search platform has rapidly captured both the advertisers and search engine companies, attributing to is high precision, low barriers to entry, and easy cost control advantages. The sponsored search technology now becomes the main revenue model of the domestic and international search engines. Under this background, this paper will focus on the empirical research by collecting the real costs and benefits data of advertisers from the sponsored search in Taobao to make up for previous lack of the research in this region and estimate the practical effect of the sponsored search. This paper will propose two new concepts:the direct and cross-category performance, to find out the reason of the cross-category performance by analyzing the different customer behavior in the sponsored search on-line shopping environment. In addition, analyze the relationship between the direct and cross-category performance and the factors which affect both of them. This paper will first bring out the literature review of the theoretical research of the sponsored search region both in and abroad to build the theoretical background of the related region. The second part of this paper will introduce the background and features of the empirical data by presenting the descriptive statistical research conclusions. The main part of this paper will be constructed by the theoretical assumptions and econometric model designing, which break the whole article into three blocks:the total click amount model; direct sale amount model and indirect sale model (the cross-category performance). The next part will show the result of regression analysis and discuss all the findings in this empirical research. Finally, based on all the findings, this paper will conclude both the theoretical recommendations and application-use suggestion, followed by the improve advices for the future research. Considering all the findings in this research, the conclusion and suggestion will be brought out both theoretically and practically:1. increasing the number of the sponsored search words will lead to better affect on the performance than raising the bid price for each sponsored search words.2. The goods attribute and business purposes should be fully considered to make the right decision to choose the sponsored search words.3. The goods which can lead to relate-purchase should be considered to promote by the sponsored search promotion, since the indirect sales amount can be raised by the direct sales performance.
在此背景下,本文中以淘宝网直通车这一竞价排名搜索栏目的数据为基础,搜集了广告主真实的成本和收益数据进行实证研究,以弥补以往研究在此方面的不足,真实地估计竞价排名的实际效果。本论文通过分析消费者在使用竞价排名搜索应用时的不同行为流程,提出了商家直接绩效和溢出绩效两个新概念,研究在竞价搜索排名推广模式中,跨类溢出绩效是怎样产生的。并建立计量模型研究最初的竞价排名搜索直接绩效与溢出绩效之间的存在关系以及影响溢出绩效和直接绩效的影响因素。
本论文内容首先将在文献综述部分介绍目前国内外对于竞价排名搜索的相关理论研究,以及本论文所涉及领域相关的理论背景。然后通过数据背景部分说明此次研究所使用数据的来源和特点,详细说明数据中各个变量的具体含义,构成和特点以及对数据进行的描述性统计分析结果发现。在之后的假设与模型设计阶段将把这个课题分为三个部分：总点击量绩效模型、直接销售数量绩效模型和间接销售数量绩效模型,展示此次研究的计量模型设计,并在之后对实证数据的回归结果进行分析。最后,根据研究结果中的结论提出针对于广告主商家在关键词和竞价策略选择方面有指导意义的管理应用建议,并对此次研究作出总结、提出需要进一步改进的方向,且.为未来更深入的研究提出建议。
本文涉及的实证研究,使用来自淘宝网直通车栏目下的时间跨度达60天1105个商家共40162条实际销售面板数据进行研究。并主要得出以下几点结论：1.商品信息总点击量直接影响商家的直接销售绩效,而直接销售绩效会对间接溢出绩效产生正相关的作用。2.商家对关键词的选择会对商家各类绩效产生不同的影响：a.关键词的长度对总点击量绩效有负相关的影响效果,但会使得直接销售绩效正向增长。b.关键词中的品牌信息,在被关联商品为具有高品牌影响度性质的条件下,会对商品信息的点击量产生正相关的影响,而对于直接和溢出间接销售绩效则有负相关的作用。c.关键词中的促销信息会对商品信息的总点击量绩效产生显著正相关的影响。3.商家的信誉级别因素会对商家的直接、溢出间接销售绩效产生重大的影响。信誉级别越高,销售绩效越好。4.在该数据来源商品特征环境下,商品的价格因素对商家各类绩效影响均不显著。
综合以上的研究结果,本文针对商家参与竞价排名搜索推广中所涉及的三个选择策略提出了以下一些启示：1.增加与商品关联的竞价关键词的数量比提升单一关键词竞拍价格更能带来高的实际收益。2.选择竞价关键词时应当推广商品的特征属性以及推广目的,选择具有不同属性特征的关键词,获得不同的绩效影响。3.在选择推广商品时,应当选择能引致关联消费的商品,通过增加直接销售绩效,进而使得间接商品的销售绩效得到提升。
Since 1970, the start of Internet applications, Internet technology has undergone a crazy non-rational stage of development. After went through the re-engineering phase, it now has entered into the far-reaching and deep evolution phase of the web2.0 implications era. In the last 40 years, the economic development of the Internet in China was overwhelming. From 2005 to 2010, the development rate of the Internet economy was much faster than domestic GDP growth rate. Internet application development has made a profound impact on the domestic economy. Especially in recent years, the web search technology has been fully developed. Currently the user amount of search engines has taken more than 70% of the total number of all Internet users. The number of search engine user has exceeded 200 million. The current instant messaging service, e-mail services and search engines have become the three major applications of the internet. Since 2000, first launched, the sponsored search technology which was carried by the web search platform has rapidly captured both the advertisers and search engine companies, attributing to is high precision, low barriers to entry, and easy cost control advantages. The sponsored search technology now becomes the main revenue model of the domestic and international search engines. Under this background, this paper will focus on the empirical research by collecting the real costs and benefits data of advertisers from the sponsored search in Taobao to make up for previous lack of the research in this region and estimate the practical effect of the sponsored search. This paper will propose two new concepts:the direct and cross-category performance, to find out the reason of the cross-category performance by analyzing the different customer behavior in the sponsored search on-line shopping environment. In addition, analyze the relationship between the direct and cross-category performance and the factors which affect both of them. This paper will first bring out the literature review of the theoretical research of the sponsored search region both in and abroad to build the theoretical background of the related region. The second part of this paper will introduce the background and features of the empirical data by presenting the descriptive statistical research conclusions. The main part of this paper will be constructed by the theoretical assumptions and econometric model designing, which break the whole article into three blocks:the total click amount model; direct sale amount model and indirect sale model (the cross-category performance). The next part will show the result of regression analysis and discuss all the findings in this empirical research. Finally, based on all the findings, this paper will conclude both the theoretical recommendations and application-use suggestion, followed by the improve advices for the future research. Considering all the findings in this research, the conclusion and suggestion will be brought out both theoretically and practically:1. increasing the number of the sponsored search words will lead to better affect on the performance than raising the bid price for each sponsored search words.2. The goods attribute and business purposes should be fully considered to make the right decision to choose the sponsored search words.3. The goods which can lead to relate-purchase should be considered to promote by the sponsored search promotion, since the indirect sales amount can be raised by the direct sales performance.
引文
[1].艾瑞咨询.2010年中国搜索引擎年度研究报告[R].2011.
[2]. Chen, J., D. Liu, and A.B.b. Whinston, Auctioning Keywords in Online search[J], Journal of Marketing (73) 2009.
[3].艾瑞咨询.2009-2010年中国搜索引擎行业发展报告[R].2010.
[4].徐涛.企业搜索引擎广告策略研究[D].上海：华东师范大学硕士论文.2010.
[5]. Edelman B, Ostrovsky M, Schwarz M. Internet advertising and the generalized second price auction[R]. US:National Bureau of Economic Research,2005: 11765.
[6]. Aggarwal G, Goe A, Motwani R. Truthful auctions for pricing search keywords [C]//Proceedings of ACM Electronic Commerce,2006.
[7]. Feng Juan, Bhargava H K. Pennock D M. Implementing sponsored search in web search engines:Computational evaluation of alternative mechanisms [J]. Informs Journal on Computing,2007,19(1):137-148.
[8]. Varian H R. Position auctions [R]. Berkeley:University of California,2005.
[9]. Brooks, N.2004. The Atlas rank report II:How search engine rank impacts conversions[R]. Atlas Institute.
[10]. Baye, M.,R.Gatti, P.Kattuman, J.Morgan.2008. Clicks, Discontinuities, and Firm Demand Online, Working paper, SSRN.
[11]. Robu V, H. Poutr'e and S. Bohte, The complex dynamics of sponsored search markets, SSRN eLibrary,2009.
[12]. Aggarwal, G., J.Feldman, S.Muthukrishnan.2006. Bidding to the top:VCG and equilibria of position-based auctions. In Proceedings of the Fourth Workshop on Approximation and Online Algorithms.
[13]. Kiritchenko S. and M. Jiline, Keywords Optimization in Sponsored Search via Feature Selection, JMLR:Workshop and Conference Proceedings 4:122-134.
[14]. Ghose, A and S. Yang, Modeling Cross-Category Purchases in Sponsored Search Advertising, SSRN eLibrary,2008.
[15]. Amer-Yahia, S., S. Lahaie, and D. M. Pennock (2008). Towards a generic bidding standard for online advertising. In EC.08:Proceedings of the 9th ACM conference on Electronic commerce.
[16].王非.网络广告关键词选择优化[J].计算计工程与应用,2009,45(7)： 227-229.
[17]. Yao, Song and Carl F. Mela (2009), Sponsored Search Auctions:Research: Opportunities in Marketing, SSRN eLibrary,2009.
[18]. Ghose, A and S. Yang, Modeling Cross-Category Purchases in Sponsored Search Advertising, SSRN eLibrary,2008.
[19]. Rutz, O.J., and R.E. Bucklin, A Model of Individual Keyword Performance in Paid Search Advertising[E], SSRN eLibrary,2007.
[20]. Ghose, A., and S. Yang, An Empirical Analysis of Search Engine Advertising: Sponsored Search in Electronic Markets[J], Management Science,2009.
[21]. Ghose, A., and S. Yang, An Empirical Analysis of Search Engine Advertising: Sponsored Search in Electronic Markets[J], Management Science,2009.
[22]. Agarwal, A., K. Hosanagar, and M.D. Smith, Location, Location, Location: An Analysis of Profitability of Position in Online Advertising Markets, SSRN eLibrary,2008.
[23]. Rutz, O.J., and R.E. Bucklin, A Model of Individual Keyword Performance in Paid Search Advertising[E], SSRN eLibrary,2007.
[24]. Ghose, A., and S. Yang, An Empirical Analysis of Search Engine Advertising: Sponsored Search in Electronic Markets[J], Management Science 2009.
[25]. Jerath K. L. Ma, Y. Park and K. Srinivasan, A "Position Paradox" in Sponsored Search Auction, SSRN E-library,2009.
[26]. Feng, J., H.K. Bhargava, and D. Pennock Implementing Paid Placement in Search Engines:Computational Evaluation of Alternative Mechanisms[J]. INFORMS Journal of Computing 2007,19(1):137-148.
[27]. Feng, J., H.K. Bhargava, and D. Pennock, Implementing Paid Placement in Search Engines:Computational Evaluation of Alternative Mechanisms[J]. INFORMS Journal of Computing 2007; 19(1):137-148.
[28]. Rutz, O.J., and R.E. Bucklin, A Model of Individual Keyword Performance in Paid Search Advertising[E], SSRN eLibrary,2007.
[29]. Katona, Z., and M. Sarvary, The race for sponsored links:Bidding patterns for search advertising[W],2008. SSRN E-library. University of California, Berkeley
[30]. 陈磊,刘奕群,茹立云.基于用户日志挖掘的搜索引擎广告效果分析[J].中文信息学报,2008,22(6)pp 92-97.
[31]. Chan, T.Y., Y. Xie, and C. Wu, Measuring the Value of Customer Acquisition from Search Advertising, SSRN eLibrary 2009.
[32]. X. Dreze, F. Hussherr, Internet Advertising:Is Anybody Watching? [J]. Journal of Interactive Marketing,2003, Vol.17 (4):8-23.
[33]. Rutz, O. J. and R. E. Bucklin (2008). From generic to branded:A model of spillover dynamics in paid search advertising. SSRN E-library.
[34]. Ghose, A and S. Yang, Modeling Cross-Category Purchases in Sponsored Search Advertising, SSRN eLibrary,2008.
[35]. 傅智辉.搜索引擎广告业务模式分析[J].高校图书情报论.2006,5(4).
[2]. Chen, J., D. Liu, and A.B.b. Whinston, Auctioning Keywords in Online search[J], Journal of Marketing (73) 2009.
[3].艾瑞咨询.2009-2010年中国搜索引擎行业发展报告[R].2010.
[4].徐涛.企业搜索引擎广告策略研究[D].上海：华东师范大学硕士论文.2010.
[5]. Edelman B, Ostrovsky M, Schwarz M. Internet advertising and the generalized second price auction[R]. US:National Bureau of Economic Research,2005: 11765.
[6]. Aggarwal G, Goe A, Motwani R. Truthful auctions for pricing search keywords [C]//Proceedings of ACM Electronic Commerce,2006.
[7]. Feng Juan, Bhargava H K. Pennock D M. Implementing sponsored search in web search engines:Computational evaluation of alternative mechanisms [J]. Informs Journal on Computing,2007,19(1):137-148.
[8]. Varian H R. Position auctions [R]. Berkeley:University of California,2005.
[9]. Brooks, N.2004. The Atlas rank report II:How search engine rank impacts conversions[R]. Atlas Institute.
[10]. Baye, M.,R.Gatti, P.Kattuman, J.Morgan.2008. Clicks, Discontinuities, and Firm Demand Online, Working paper, SSRN.
[11]. Robu V, H. Poutr'e and S. Bohte, The complex dynamics of sponsored search markets, SSRN eLibrary,2009.
[12]. Aggarwal, G., J.Feldman, S.Muthukrishnan.2006. Bidding to the top:VCG and equilibria of position-based auctions. In Proceedings of the Fourth Workshop on Approximation and Online Algorithms.
[13]. Kiritchenko S. and M. Jiline, Keywords Optimization in Sponsored Search via Feature Selection, JMLR:Workshop and Conference Proceedings 4:122-134.
[14]. Ghose, A and S. Yang, Modeling Cross-Category Purchases in Sponsored Search Advertising, SSRN eLibrary,2008.
[15]. Amer-Yahia, S., S. Lahaie, and D. M. Pennock (2008). Towards a generic bidding standard for online advertising. In EC.08:Proceedings of the 9th ACM conference on Electronic commerce.
[16].王非.网络广告关键词选择优化[J].计算计工程与应用,2009,45(7)： 227-229.
[17]. Yao, Song and Carl F. Mela (2009), Sponsored Search Auctions:Research: Opportunities in Marketing, SSRN eLibrary,2009.
[18]. Ghose, A and S. Yang, Modeling Cross-Category Purchases in Sponsored Search Advertising, SSRN eLibrary,2008.
[19]. Rutz, O.J., and R.E. Bucklin, A Model of Individual Keyword Performance in Paid Search Advertising[E], SSRN eLibrary,2007.
[20]. Ghose, A., and S. Yang, An Empirical Analysis of Search Engine Advertising: Sponsored Search in Electronic Markets[J], Management Science,2009.
[21]. Ghose, A., and S. Yang, An Empirical Analysis of Search Engine Advertising: Sponsored Search in Electronic Markets[J], Management Science,2009.
[22]. Agarwal, A., K. Hosanagar, and M.D. Smith, Location, Location, Location: An Analysis of Profitability of Position in Online Advertising Markets, SSRN eLibrary,2008.
[23]. Rutz, O.J., and R.E. Bucklin, A Model of Individual Keyword Performance in Paid Search Advertising[E], SSRN eLibrary,2007.
[24]. Ghose, A., and S. Yang, An Empirical Analysis of Search Engine Advertising: Sponsored Search in Electronic Markets[J], Management Science 2009.
[25]. Jerath K. L. Ma, Y. Park and K. Srinivasan, A "Position Paradox" in Sponsored Search Auction, SSRN E-library,2009.
[26]. Feng, J., H.K. Bhargava, and D. Pennock Implementing Paid Placement in Search Engines:Computational Evaluation of Alternative Mechanisms[J]. INFORMS Journal of Computing 2007,19(1):137-148.
[27]. Feng, J., H.K. Bhargava, and D. Pennock, Implementing Paid Placement in Search Engines:Computational Evaluation of Alternative Mechanisms[J]. INFORMS Journal of Computing 2007; 19(1):137-148.
[28]. Rutz, O.J., and R.E. Bucklin, A Model of Individual Keyword Performance in Paid Search Advertising[E], SSRN eLibrary,2007.
[29]. Katona, Z., and M. Sarvary, The race for sponsored links:Bidding patterns for search advertising[W],2008. SSRN E-library. University of California, Berkeley
[30]. 陈磊,刘奕群,茹立云.基于用户日志挖掘的搜索引擎广告效果分析[J].中文信息学报,2008,22(6)pp 92-97.
[31]. Chan, T.Y., Y. Xie, and C. Wu, Measuring the Value of Customer Acquisition from Search Advertising, SSRN eLibrary 2009.
[32]. X. Dreze, F. Hussherr, Internet Advertising:Is Anybody Watching? [J]. Journal of Interactive Marketing,2003, Vol.17 (4):8-23.
[33]. Rutz, O. J. and R. E. Bucklin (2008). From generic to branded:A model of spillover dynamics in paid search advertising. SSRN E-library.
[34]. Ghose, A and S. Yang, Modeling Cross-Category Purchases in Sponsored Search Advertising, SSRN eLibrary,2008.
[35]. 傅智辉.搜索引擎广告业务模式分析[J].高校图书情报论.2006,5(4).