[雙語翻譯]--外文翻譯--理解在線產(chǎn)品評級顧客滿意度模型（原文）

1、Understanding online product ratings: A customer satisfaction modelTobias H. Engler n, Patrick Winter, Michael SchulzPhilipps University at Marburg, School of Business Lin et al., 2011; Park et al., 2007). Onthe other h

2、and, online retailers and manufacturers increasinglyrely on customer feedback to enrich their marketing strategy(Chen and Xie, 2008; Cui et al. 2012), to adjust product listings(e.g. via relevance sorting), and to create

3、 additional revenuestreams (Mudambi and Schuff, 2010). For these reasons, it is notsurprising that nearly all major online retailers such as Amazon.com or Ebay.com have implemented product rating functionalities. Researc

4、hers, mainly from the fields of marketing and in-formation systems, have adopted the topic and not only started tostudy the effects of online product ratings (e.g., on sales) but alsotheir nature and determining factors.

5、 A common assumption ofprior studies in the latter stream is that the baseline of a product'sonline ratings reflects its true quality. Various biases such as socialdynamics or cultural influences were introduced to a

6、ccount for theunexplained part of the variance. However, empirical evidencesuggests that online ratings do not accurately reflect a product'strue quality (e.g., Hu et al., 2006; Koh et al., 2010). Since the in-fluenc

7、e of ratings on sales remains unaffected, retailers are left inan uncomfortable situation: it is difficult for them to adjust mar-keting strategies on the basis of online product ratings withoutknowing what they actually

8、 reflect. Hence, the objective of this study is to find out what reallybuilds the baseline of online product ratings and thereby refinetheir current interpretation. We argue that the weak explanatorypower of product qual

9、ity for online reviews is not only caused byactual biases: it is mainly caused by product ratings reflectingcustomer satisfaction than being a valid measure for productquality. This approach does not solely rely on produ

10、ct quality asthe baseline for the rating but also integrates the customer's ex-pectation of the product in the pre-purchase phase. Correspond-ingly, we present a customer satisfaction model of online productratings b

11、ased on the considerations of Fornell (1992) and West-brook and Reilly (1983). We model the customer's pre-purchaseexpectation of the product and the actual performance as pre-dictors of online ratings using structur

12、ed equations. We validateour model by applying it to two datasets (digital cameras andtelevisions) collected from the German website of Amazon.com.The results indicate that both a customer's expectation of a pro-duct

13、 and the actual performance significantly influence the ratingscustomers assign to a product, supporting the proposed inter-pretation of online product ratings. Several other observations in the datasets can help to get

14、amore comprehensive view of online product ratings and are worthmentioning. First, we find that online ratings carry some percen-tage of unobservable information that cannot be predicted (usingmetrics from the website).

15、Second, the data shows indications forconfirmation, acquisition, and under-reporting biases.Contents lists available at ScienceDirectjournal homepage: www.elsevier.com/locate/jretconserJournal of Retailing and Consumer S

16、erviceshttp://dx.doi.org/10.1016/j.jretconser.2015.07.0100969-6989/ Bolton, 1998).It describes the order of belief updating over time as a process ofanchoring and adjustments. The central message of the belief-adjustment

17、 model is that individuals do not directly react to a newstimulus but rather adjust their prior expectations on the specifictopic to the new stimulus while sustaining in the vicinity of theoriginal anchor (cf. Oliver, 19

18、80). Thus, pre-purchase expectationsshould have a positive impact on satisfaction. It was found to beapplicable in various contexts. This leads us to assume that thisprocess also takes place in the context of online shop

19、ping and thepre-purchase evaluation of products. First, customers form anexpectation what the product might be like on the basis of in-formation found on the product website. In a second step, theyadjust this anchor with

20、in a reference frame set by the initial jud-gement when being confronted with the product's performanceafter the purchase and delivery. Hence, we hypothesize:Hypothesis H1. : Pre-purchase expectations (EXP) have a po

21、sitiveimpact on the score of online product ratings (PRO).The direct effect of performance on satisfaction is supported bythe value-percept disparity model developed by Westbrook andReilly, (1983). They posit that satisf

22、action is a general perceptionbased on the evaluation of customers' experiences with a product.A high satisfaction can, therefore, only be achieved if a product isable to fulfill the customer's needs. This mechan

23、ism is consistentwith findings from Churchill and Suprenant (1982). The results oftheir study suggest that satisfaction with a durable good can bepredicted by the product performance to a considerable extent.Further stud

24、ies also support this direct effect of performance onsatisfaction (Anderson and Sullivan, 1993; Fornell, 1992). Trans-ferred to the online environment, this means that online productratings are indeed influenced by the e

25、xperienced quality of theproduct, as assumed by prior research (e.g., Koh et al., 2010). Theproduct's performance should, therefore, have a positive effect onthe score of online ratings. Thus,Hypothesis H2. : A produ

26、ct's post-purchase performance (PER)has a positive impact on the score of online product ratings (PRO).4. Research method and data analysis4.1. Measurement and data collectionThe research model was tested using crawl

27、ed data of camerasand televisions to address the two major shortcomings of priorresearch as described above. Books and movies can be classified asexperience goods while cameras and televisions are search goods(cf. Nelson

28、, 1970,, 1974). The ratings of experience goods heavilydepend on personal feelings, cannot be evaluated on the basis ofspecific characteristics, and may vary across different individuals(Mudambi and Schuff, 2010; Weather

29、s et al., 2007). Whereas thebeauty of a book or a movie is in the eye of the beholder, it ispointless to argue about objective measures such as battery life-time or viewing angel stability. Search goods such as cameras a

30、ndtelevisions can be evaluated using a more systematic approach(Cui et al., 2012) including rather objective criteria such as tech-nical functions (e.g., megapixels) into the evaluation process,hence, increasing rating r

31、eliability.4.1.1. Expectation The aim of this research is to identify factors that constitute thescore of online ratings made by customers of an online shop. Forthis, we adopted the customer's perspective and focus o

32、n quan-titative data that can be included in the evaluation by quicklyoverlooking the product's description on the website (see Fig. 3).Accordingly, expectation was captured using three indicators thatcan be evaluate

33、d by customers this way before buying the pro-duct: the average score of previous ratings, the product price, andbrand reputation. While the score of previous ratings is the majorsource of information for online customer

34、s (Koh et al., 2010; CuiFig. 2. Consumer satisfaction model of online product ratings.Fig. 3. Product description on amazon.com and measurement model ofexpectation.T.H. Engler et al. / Journal of Retailing and Consumer S