The current chapter introduces CUSENT, a tool for semantics-enhanced sentiment analysis of customer opinions expressed in corporate blogs. The research work presents the examination of emotions and sentiments from the perspective of information systems, and, in particular, provides a review of the principal efforts for the conceptualization of emotions and sentiments in texts. Subsequently, a description of the proposed architecture of the platform is outlined. The authors aim to contribute a solution which automates the analysis of customer opinions in company blogs that relies on existing techniques, but further exploits these methods to store and reuse customer feedback. The novel combination of opinion mining with an ontology of emotions can thus be used in organizational creation and innovation processes, which characterize the new forms of communication derived from the institutional and commercial use of Web 2.0.
TopIntroduction
The Internet is changing the forms in which people relate among each other. Many of the tools which are unique to the Internet, such as blogs, wikis or virtual worlds, are used by people as the fundamental tools to construct relationships. According to O’Reilly (2005) a fundamental principle of Web 2.0 is that users add value by generating content through these applications, resulting in network effects among the community of users. This circumstance is instigating more and more organizations to attempt to exploit the gains of Web 2.0. In this new scenario, collaboration and co-creation with customers has emerged through habitual use of Web 2.0, by both companies and clients alike (Vargo & Lusch, 2004). According to Manyika, Roberts & Sprague (2007), in the environment of Information Technology related technology trends, co-creation, alongside the use of customers as innovators and making businesses from information, has been termed as one of the trends to watch. It is precisely these three trends which comprise the key motivation for the current research.
In the first place, in relation to distributed co-creation, today many companies involve customers and other players in the creation of new products. Current technology enables companies to benefit from partial externalization of the co-creation process. In the current research, the aim is to extend the process of co-creation, delegating the control to customers, as illustrated in the work by Prahalad & Krishnan (2008). The work aims to benefit from one of the most important assets held by these fundamental stakeholders: their opinion oriented towards co-creation.
In the second place, it is aimed to exploit the innovation capacities of customers. Involving customers in design, testing and the after-sales process can be used as an element to gain better insights into customer needs and behavior, and may be able to cut the cost of acquiring customers, engender greater loyalty, and speed up development cycles (Manyika, Roberts & Sprague, 2007). By means of the application of technologies characteristic of the Web 2.0, a further objective of the research is to benefit from the metadata generated by such a platform of interaction and communication, dealing with some of its weaknesses, but at the same time benefiting from its undeniable advantages.
A final objective can be defined as the following: Making businesses from information. More concretely, the task is to exploit the interactions produced in Web 2.0 in order to generate business, using the capacities for processing and analysis of information provided by the technology. According to a study by McKinsey consultants (McKinsey, 2007) where 2,847 executives were interviewed, respondents inform that Web 2.0 technologies are strategic and that they plan to increase these investments, moreover, they say they are using Web 2.0 technologies to communicate with customers and business partners and to encourage collaboration inside the company. More precisely, executives inform blogs are also frequently mentioned as a channel to communicate with customers and, in some cases, critics.
This is the acting environment of CUSENT. Applying text analysis methods, an analysis has been carried out of the responses which clients post on a determined organizational blog. The use of semantics enables the categorization of sentiments in an ontology which is populated with emotions, which CUSENT detects from the results of a text classification algorithm constructed to perform sentiment analysis. The subsequent use of the information extracted aids innovation processes by means of the active support of customers.
The remainder of the chapter has been structured as follows. Initially, the basic concepts of sentiments, emotions and semantics are outlined. Secondly, the principal initiatives for sentiment analysis are discussed. The next section sketches how the CUSENT architecture and implementation fits in the picture. Lastly, the principal conclusions of the paper are presented, and some suggestions for resolving the problems encountered are outlined, as well as proposals for future research.