Paintings & Photography

An integrated framework For modelling the adoptive behaviour of online product recommendations

An integrated framework For modelling the adoptive behaviour of online product recommendations
of 5
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  An Integrated Framework For Modelling The Adoptive Behaviour Of Online Product Recommendations Wen-Shan Lin  National Chiayi University, Taiwan  E-mail: Abstract  A new-product recommendation system supports users in making buying decisions. The method representing recommending results in accordance with users’ preferences is one of the crucial technologies for persuading users to adopt the recommendations. A body of literature proposes technical improvements in recommendation systems (RS). However, the research that investigates users’ behaviour toward adopting results presented by the  RS is scarce.This paper targets the online scenario that customers provide post-purchase opinions; in this case, e-word-of-mouth   eWOMs). It investigatesthe choice model in terms of adoptive and evaluative behaviourof the eWOMs receivers (users of the recommendation systems). This paper presents ongoingresearch that proposes an integrated  framework that links literature onthe use ofproduct recommendation systems. 1. Introduction The main functions of product recommendation systems are to filter information and to present results.With respect to information-filtering technologies, content-based technology and collaborative technology are the dominant ones. Content-based recommendation mechanism based on users’ preferences has significantly improved efficiency [18]. At the same time, the used methods of results presentations and explanations are revealed have links with the persuasive values perceived by consumers [7]. However, there is no integrated framework proposed in the literature by integrating trust factors and persuasive values denoted byrecommendation results. Therefore, this paper introduces these two factors in investigating users’ adoptive behaviour in adopting new product recommendations.This paper targets the online scenario in which customers provide post-purchase opinions: e-word-of-mouth  eWOMs  in this case. It investigatesthe discrete choice model in terms of adoptive and evaluative behaviour of eWOMs receivers (users of the recommendation systems). The research objectives are: (1) t o empirically investigate users’ choice model in adopting product recommendations, (2) t o examine users’ preferences in adopting product recommendations based on geographical information,(3) to inspect the choice model for adopting product recommendations,and (4) to propose an integrated model for product recommendation systems.A literature review is presented in section 2. The research methodology is specified in section 3. Preliminary data analyses are presented in section 4. Discussions and conclusions are given in section 5. 2. Relates Work A recommendation system assists users in making decisions in meeting their requirements. In the Internet era, online product recommendations are developed to assist online shoppers in making buying decisions. It solves the problem of information overload. The major recommendation mechanisms are collaborative filtering, content-based filtering, and knowledge-based recommender applications [1].In general, the recommendation mechanisms applied in the domain of electronic commerce have links with personalisation and self-machine learning. By using the related mechanisms, it provides product recommendations based on matching product  preferences with the buying choices made by other users. It also  plays an important role in consumers’  buying decisions [9]. 2.1 Users of recommendation systems An online recommendation system is assigned as intelligent shopping agents, assists users in solving the problems of information overload by providing solutions that meet users’ needs  [14, 15]. Arecommendation system is an important mechanism in mediating computer and users by providing useful search results based on individualised preferences [6]. However, the persuasive value denoted by recommended results and the link between the developments of recommendation systems has not Copyright © 2009 by the Institute of Electrical and Electronics Engineers, Inc. All rights reserved.   been not fully covered in the literature. Furthermore, electronic word-of-mouth, as messages on products  by online users, has grown in popularity [12]. The  posted messages are transformed into neutral and trustworthy information that have ahigh impact onconsumer behaviour. Therefore, the case of eWOM is adopted in this paper to explore users’ discrete choice model in the domain of recommendation applications. The proposed research model  Figure 1  and hypotheses are detailed below.Theserecommendation systems are expected to assist users in solving problems [14]. Besides, as the trust factors and the persuasion value conveyed by the recommendation systems and their results  presentations, the research model is divided twofold: (1) the interactive process between users and recommendation systems,and (2) persuasive value denoted by the explanations of recommendations. Theindependent variables in each stagewill be described.   Figure 1. Research Model 2.2 Persuasion through the interactive process Recommender systems can both persuade and recommend. Three cues in the preference-elicitation  process can influence users’ perceptions of how well the recommendation matches their preferences: (1) relevance, (2) transparency, and (3) effort [6]. Thus, the following hypothesis is proposed: H1 a  Topic-relevance increases the perceived value and trust while interacting with recommendation systemsH1  b : Transparency increases the perceived value and trust while interacting with recommendation systems H1 c : Greater efforts make the interactive process  between interacting with recommendation systems less enjoyable 2.3 Task-technology fit model The task-technology fit model  TTF  is applied to test user  s’  behaviourin adopting information systems [3].The TTF model originatedfrom the technology acceptance model  TAM   that  proposes task requirements and tool functionality asthe two main factors in measuring actual tool use and individual performance with respect to the user side [5, 11, 6]. TTF can therefore  be applied to test users’  perceived value in the interactive process of adopting the recommendation systems in this paper. Thus, the following hypothesis is proposed:H2 a  Perceived individualisation hassignificant effect on affecting user discrete choice behaviour in adopting recommendation systemsH2  b : Familiarity has significant effect on affecting user discrete choice behaviour in adopting recommendation systems TransparencyRelevancyH2H2  b H6H4H3H1 c H2 a H1  b H1 a Perceived value of  processPerceived individualizationFamiliarityPerceived enjoyment of  process Source credibility of eWOM H5   User discrete choice  behaviour of adopting recommendation systemEffortThe interactive process between users and recommendation systemsPersuasive value denoted  by the explanations of recommendationsProduct involvementHerd value  2.4 Perceived enjoyment process  General users are not certain about their  preferences; therefore, if questions are raised by the recommendation systems that clarify users’  preferences, users will be more satisfied with those systems [6].Hedonic value is a cognitive factor that affects users ’  adoption of electronic commerce or online services, such as taking online psychological-test to identify the personal characteristics that can raise the level of hedonic value. Thus, the following hypothesis is proposed: H3: Hedonic value perceived by users has significant effect on user  s’  behaviour in adopting recommendation systems 2.5 Product involvement and perceived herd value Product involvement influences users in accepting  purchase recommendations [15]. With low-involvement products, users tend to follow the  buying recommendations made by the majority of other buyers; users also tend to follow buying recommendations that have high quality in terms of full explanations for buying high-involvement  products. Thus, the following hypothesis is proposed: H4  Product involvement has an effect on the user choice modelin adopting recommendationsAcollaborative recommendation is one of the most popular mechanisms for revealing search results by recommendation systems [4]. In the domain of electronic commerce, this mechanism has several applications, e.g. collaborative recommendation extracts product information that are the most popular.By pushing information that has good popularity can persuade shoppers that have similar needs. In other words, this mechanism highlights the importance of herd value by matching users who have similar search requirements in order to improve the efficiency of search and recommendation [13]. The importance of herd value in the network knowledge exchange is stated in the social exchange theory [2].People tend to follow the comments made by the majority of other users in order to decrease the risk and obtain social acceptance. Thus, the following hypothesis is  proposed: H5: Herd value has an effect on user  s’  discrete choice behaviour in adopting recommendations 2.6 Source credibility theory Opinion-seeking behaviour is the main motivation of recommendation-searching users [10]. Opinion leaders are those whose views are most influential for people with similar backgrounds or interests. Their opinions are classified as  1   product news word-of-mouth, (2) advice giving word-of-mouth, (3) personal experience word-of-mouth[10].The source credibility theory proposed by Hovland and Weiss [8]positsout word-of-mouth originatesfrom different channels or that different subjects have different influences on receivers. In other words, the opinions ofpeople withfull experience of aproduct have a greatinfluence on others. Thus, the following hypothesis is proposed:H6: Source credibility of eWOM has affects user  s’ discrete choice behaviour in adopting recommendations 3. Research Methods The target respondents were online users with experiences in online recommendation systems.A university campus is selected as the location to approach qualified samples and distribute the questionnaire. In this stage, 50 subjects were randomly selected to complete a research questionnaire. 3.1 Measurement development Thirty-six variables are created and included in the questionnaire based on the research model and related literature as presented in section 2. The questionnaire comprised two parts, namely variable measurement and personal information sections. The main part of the questionnaire included variable measurements that are derived from six constructs: (1) persuasion through the interactive process, (2)  perceived value and trust based on task-technology fit model, (3) perceived enjoyment process, (4)  product involvement, (5) perceived herd value and (6) trust level of eWOM based on source credibility theory. Personal information requested in this study included gender, age, education background,  personal allowance, marriage, occupation, experiences in utilisingsearchengine/ recommendation systems, online shopping experiences, average online paymentand bought  products.This preliminary data collection was conducted in Taiwan and thus the measurement scale was  presented in Chinese. To ensure the content validity, two postgraduate students and one expert in information technology have reviewed the research instruments. The survey questionnaire was then  pilot-tested by involving 50 samples in order to identify any areas requiring modification. The pilot test was designed to eliminate any ambiguous sentences contained in the questionnaire. The questionnaire was then modified based on their suggestions. Table 1 demonstrates the measurement construct and numbers of tested constructs with the denoted code.  Table 1. Variable definitions and measurements CodeMeasurement Construct No. of tested variablesSourcesROERelevancy, transparency and efforts required10[6]VTPerceived value and trust10[5,11,6]HEHedonic value4[6]PVPersuasive value12[8,13,2] 4.Data analysis and results Fifty questionnaires were randomly selected at a University campus because young users who have at least a college background tend to be heavy online users. Forty-five valid questionnaires were returned,yielding a return rate of90%. All subjects were given sufficient time  10 minutes  to complete the questionnaire. Among the 45usable questionnaires, all subjectshave online shopping experiences and have utilisedrecommendation systems. Data from these 45online shoppers were used to analyse the research model.Among thesubjects,71.1%were male and 28.9%were female;their average age was 24years old. All of the subjectshaveattended university and 11%werepostgraduate students. 4.1 Content validity This preliminary study consists of relatively small samples. And the questionnaire is developed based on literature; therefore, the content validity should meet the requirements. 4.2Instrument reliability Cronbach’s α  is employed to test instrument reliability. As it is suggested that the Cronbach’s alpha value ranged above 0.8 deemed the high level of reliability. Results indicate that all the values are reasonably acceptable  Cronbach’s α =0.934  4.3Preliminary results The t-test is applied to test the significant level of each variable perceived by participants   N=45  .The preliminary results reveal that mosttested variables indicate statistically significant   p<.005  and been perceived statistically important  mean>3  .There is one tested variable  PV3  grouped to the construct of perceived persuasion needs to exclude in the final questionnaire  mean=2.87, p>0.005  . 5.Conclusions This paper introduces concepts of trust and  persuasive value in investigating users ’  adoptive  behaviour toward adopting recommendation systems. The task-technology fit model and source-credibility theory are both applied in the framework for further discovering users ’  choice behaviour. Based on the  preliminary results, the pilot-tested questionnaire  based on the proposed research model requires slight amendments for preceding an empirical test in the next stage. 6. References [1] H.J. Ahnm, “Utilizing popularity characteristics for  product recommendation”,  International Journal of  Electronic Commerce ,2006,11(2), 59-80. [2] Y.F. Chen, “ Herd behavior in purchasing books online ” , Computers in Human Behavior  ,2008,24  5  , 1977-1992.[3] M.T. Dishaw and D.M. Strong, ”Extending the technology acceptance model with task-technology fit constructs”,  Information & Management  , 1999,36, 9-21. [4] A. Felfernig and B. Gula, “ An empirical study on consumer behavior in the interaction with knowledge- based recommender applicat ions”,  Proceedings of the 8 th  IEEE International Conference on E-Commerce Technology and the 3 rd   IEEE International Conference on  Enterprise Computing, E-commerce, and E-service ,2006, 288-296. [5] S.Gefen, D.E. Karahanna and D. Straub, “ Trust and TAM in online shopping: An integrated Model”,  MIS Quarterly,  2003, 27(1), pp. 51-90.  [6] U. Gretzel and D.R. Fesenmaier, “ Persuasion in Recommender Systems”,  International Journal of  Electronic Commerce ,2006-2007, 11(2), 81-100. [7] S.L. Huang and F.R. Lin, “ The design and evaluation of an intelligent sales agent for online persuasion and negotiation”,  Electronic Commerce Research and  Applications ,2007, 6, 285-296. [8] C,I. Hovland and W. Weiss, “ The influence of source credibility on communi cation effectiveness”, The Public Opinion Quarterly ,1951-1952, 15(4), 635-650. [9] J.K. Kim, Y.H. Cho, W.J. Kim, J.R. Kim, and J.H. Suh, “ A personalized recommendation procedure for Internet shopping support”,  Electronic Commerce Research and  Applications ,2002, Vol.1, 301-313. [10] M.L. Richins and T. Root-Shaffer, “ The role of involvement and opinion leadership in consumer word-of- mouth: an implicit model made explicit”,  Advances in Consumer Research ,1988, 15, 32-36.[11] S.Komiak and I. Bendasat, “ The effects of  personalization and familiarity on trust and adoption of recommendation agents”,  MIS Quarterly ,2006, 30(4), 941-960. [12] J. Lee, D.H. Park, and I. Han, “ The effect of negative online consumer reviews on product attitude: An information processing view”,  Electronic Commerce  Research and Application ,2008, 7  3  , 341-352.[13] X. Li, “Informational cascades in IT adoption”, Communications of The ACM  ,2004, 47(4), 93-97. [14] W.S. Lin, “ A study of agent-based shopping support: A case study of outbound group package-tour-products in Taiwan”,  Agent Systems in Electronic Business , Li, E.Y. and Yuan, S.T., Information Science Reference, Hershey,  New York, 2008,pp.38-57. [15] D.H. Park, J. Lee. and I. Han, “The effect of on -line consumer reviews on consumer purchasing intention: The moderating role of involvement”,  International Journal of  Electronic Commerce ,2007, 11(4), 125-148.
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!