Abstract: Lead users have the characteristics of being ahead of market trend and high profitability expectation. On the one hand, they have the novel demand that will be popularized in the target market in a few years or months; on the other hand, they can obtain great benefits from the solution that meets the novel demand. With the development of information technology, lead users play an important role not only in traditional environments but also in open innovation community. Nowadays, lead users have become a powerful enabler of product innovation. However, identifying lead users online is a new topic, and related research is scarce. According to previous research, methods to identify lead users in the era of Internet include virtual stock, online forum survey and netnography. Among them, virtual stock adopts experimental methods, so it is difficult to be applied to online communities; online forum survey distributes questionnaires to community users, but the rate of responses is usually low. Relatively speaking, netnography is a more objective and feasible method of identification.
Netnography is composed of Internet and ethnography. It adopts techniques used in ethnographic research and the observation method to study online communities. It follows four steps, the online community selecting, data collecting and analyzing, reliability testing, and research ethics following. It can analyze user behaviors in online communities, and its advantages are fast, simple and inexpensive. But when netnography is used to identify lead users, there come some defects. First, it only pays attention to users with high activity who post more while ignores users with low quality. Second, it is only for food industry, not concerning about high-tech industry. Third, it does not classify lead user characteristics. Last, testing the effectiveness of extended netnography method only use the questionnaire method, lacking a more objective method. Therefore, Belz & Baumbach (2010) suggested that future research should use an extended netnography methods to solve these problems. The extended netnography method is aimed to (1) improve the original netnography method and to study various user behaviors in online community; (2) pay attention to both the users who post more and post less at the same time when selecting research objects; (3) become suitable for analysis in the research fields of consumer goods as well as high-tech products; (4) not only use qualitative methods, but also combine quantitative methods, which classifies users by factor analysis and cluster analysis, and uses the grounded theory to test effectiveness of the identification method.
This paper studies the lead user characteristics in the open innovation community at first, and proposes three characteristics, which are demand leading, active expression and community power. Demand leading refers to the demand for products from lead users before the birth of new products. These advanced demands help to realize new products, indicating that these users have the ability to innovate. It reflects quality of participation, including the ability to grasp market trends, dissatisfaction with existing products, relevant product knowledge and experience. Active expression refers to the initiative of knowledge contribution and the enthusiasm of answering questions or helping other users, indicating the great enthusiasm of user participation. It reflects the quantity of participation characteristics, including the number of questions and answers from users. Community power refers to the ability to radiate and influence other members in the community, and reflects the characteristics of interaction effects, that is, it can attract the attention of other users, and be responded by other users.
Then, based on the extended netnography method, qualitative netnography and quantitative data analysis are combined to identify lead users. First, according to Kozinets’ (2002) criteria, we selected an open innovation community, in which community members can conduct in-depth and full discussion on a topic. Then, we paid attention to both the active and inactive users in the community, collected relevant data, and identified lead users by the factor analysis and cluster analysis. At last, the grounded theory was used to sort out and compare the posting content of lead users in community and market trends, and to test the external validity. In this study, we took 103 users of the community as research object following the community selection principle. Based on Java in MyEclipse 2014, we took the web crawler method to obtain users’ data, including 7 indicators such as the user’s most valued posts, numbers of entries extracted from lead-users database, number of questions, number of answers, number of fans, number of praises, and number of other users’ responses. We standardized these data, and performed factor analysis and cluster analysis. At last, we identified 2 lead users among them.
By taking grounded theory, we used theoretical sampling, open coding and selection coding to analyze and compare the content from lead users and the new product functions on the market, to verify the external validity of the identification method. In this study, after comparing the post contents and the new functions of sweeping robots in the future market, we found that most of the suggestions put forward by lead users will be reflected in the future new product functions, which proves the effectiveness of the extended netnography method adopted to identify lead user. 
Based on previous research, this paper further expands the research on lead user identification in the open innovation community, echoes Belz & Baumbach’s (2010) suggestion that future research should adopt the extended netnography method. This study clarifies the method in selecting online community guidelines, and follows research ethics and steps consistent with the original netnography method. Moreover, it highlights that the extended netnography method is superiority to the original netnography method in community and user selection, data collection, data analysis, etc. The research helps to improve the lead user identification method and lays a solid foundation for future behavioral research on lead users. In terms of practical value, this research helps companies to identify lead users from massive online user data in this big data era with an improved online approach that is easier, more efficient, and cost-effective to fully exploit its value.

Key words: open innovation community, lead users, identification, extended netnography