Study of the long tail formation within an ewom community. The case of ciao UK
|Author/s||Olmedilla Fernández, María|
|Director||Martínez Torres, María del Rocío|
|Department||Universidad de Sevilla. Departamento de Administración de Empresas y Comercialización e Investigación de Mercados (Marketing)|
|Awards||Premio Extraordinario de Doctorado US|
|Abstract||Continuous communication among people and ubiquitous online access are fundamental characteristics of online eWOM communities that are facilitating the distribution of a broad range of products and services. eWOM communities ...
Continuous communication among people and ubiquitous online access are fundamental characteristics of online eWOM communities that are facilitating the distribution of a broad range of products and services. eWOM communities have emerged to influence customers directly and create interest with efficacy and flexibility in spite of geographic boundaries (Duan, Gu and Whinston 2008). They provide rich and objective product information that is influencing customers’ decision making (Gu, Tang and Whinston 2013, Kim and Gupta 2009, Zacharia, Moukas and Maes 2000), due to the credibility, empathy and relevance they offer to customers as opposed to the information provided by marketer-designed websites (Bickart and Schindler 2001). Through eWOM, users can freely post their reviews about any product or service, and share those reviews with other users in order to better understand a product (Hennig-Thurau, et al. 2004). Thus, through eWOM communities, a great audience of users is able to acquire knowledge from reviews concerning products and services that are less popular to the majority. In that respect, the distribution of product sales is changing due to the increment of product information available to consumers (Brynjolfsson, Hu, & Smith, 2010) facilitating the long tail phenomenon (Anderson 2004). Many authors have given a good understanding of the main idea behind long tail within sales distributions in product markets such as Amazon (Brynjolfsson, Hu, & Smith, 2003; Brynjolfsson, Hu, & Smith, 2010). However, this Thesis goes beyond and applies new methodologies –elbow criterion– and extends others –power-law distribution– by Clauset, Shalizi and Newman (2009) to mathematically measure the long tail in other environments, such as the eWOM community Ciao. Whereas most eWOM studies focus just on the potential of eWOM facilitating the long tail effect to find rare or niche products (Hennig-Thurau, Gwinner, Walsh, & Gremler, 2004; Khammash & Griffiths, 2011) and how eWOM is enabling zero-cost dissemination of information about products (Odić, Tkalčič, Tasič, & Košir, 2013) and so forth, not many noticed that for each product type enclosed in the tail of the sales distribution there might be different impacts. In this regard, the results within this Thesis might indicate that vendors could adopt alternative product strategies depending on with which niche product type (search or experience good) the tails of sales distribution would be formed. More specifically, this Thesis proposes an approach for detecting whether there is a long tail for each product type and thus, cases should be differentiated when niche products represent a significant portion of overall product sales. Likewise, given the volume of the user-generated content in the web and its speed of change this Thesis also presents two important highlights in this regard. First, the implementation of an effective web crawler that can gather and identify big amounts of user-generated content. Second, the stages followed on this crawling process, which are the identification and collection of important data, and the maintenance of the gathered data. Consequently, social science needs to develop adequate methodologies to deal with huge amounts of data, such as the one outlined within this Thesis and overcome the distance between technology and social sciences. The chosen methodology within this Thesis has been to triangulate the method of power-law distribution of data gathered with other method, the elbow criterion in order to identify the long tail. That is, to compare the all the type of products among the eWOM Ciao UK, the probability power-law distribution function was represented as a tool to measure the long tail. Besides, to extra validate such method the elbow criterion was also used to identify where was located the optimal cut-off point that distinguishes the products characterized by the long tail. Furthermore, this Thesis outlines an architectural framework and methodology to gather user-generated data the eWOM community Ciao UK. To that end, a new methodology describes the implementation of a web crawler from other disciplinary perspective: the computing science discipline. Interestingly, the present thesis aims to contribute to the study of the long tail phenomenon in an eWOM community and what product types are enclosed there. To this end the three following hypotheses where contrasted: H1: The experience products from the distribution of product categories within an eWOM are more likely to exhibit a long tail. H2: The search products from the distribution of product categories within an eWOM are less likely to exhibit a long tail. H3: The distribution of product categories within an eWOM that have high frequency events or super-hits in the short head are not particularly associated with search or experience products. The results supported all the three proposed hypotheses. In this sense, this Thesis presents important new findings. Firstly, it is evidenced that products having a long tail are those with subjective evaluation standards, which are classified as experience products. Secondly, it is also corroborated that search products, which have a high level of objective attributes in the total product assessment do not encourage the long tail phenomenon. Thirdly, there is a combination of products when there are super-hits in the short head of the distribution. Thus, those are not particularly associated with search or experience products since they contain either objective or subjective evaluation standards. Finally, it is also remarkable to highlight that not all the categories fitting a power-law distribution are characterized by a long tail and on the contrary, some of those having a long tail do not fit a power-law. In general, the findings also suggest the potentials of eWOM, which, in general, might generate a long tail effect, where a large number of small-volume vendors coexist with a few high-volume ones. Furthermore, this Thesis has contributed to both theory and practice, essentially, in three different ways: (1) with a methodology of collection of online user-generated data in the context of social sciences; (2) with the development of two more accurate methods to identify niche products within an eWOM community, providing a deeper understanding of the long tail phenomena and the type of products; and (3) with publications of refereed journals papers (indexed in JCR/JSCR) as well as conference papers related to the main topic of this Thesis.
|Citation||Olmedilla Fernandez, M. (2017). Study of the long tail formation within an ewom community. The case of ciao UK. (Tesis Doctoral Inédita). Universidad de Sevilla, Sevilla.|