Today, we live in the world with the flood of incoming and as yet unprocessed data. The Internet technology, more precisely the Information Technology (IT) is changing almost every corner of our society and life from the end of 20th century. People now connect and react to the world more easily with their extended sensors, e.g. mobile phones. As a result, the data people generate have become more prevalent, more various, and more accessible. Judging by the rate of the current technological evolution, the phenomenon will continue and spread more rapidly. An important implication of such IT invasion is that we have new opportunities to better understand humankind and our society. From the perspective of producers or firms in economic systems, people are consumers as well as value co-creators. Firms, by nature, try to understand consumers to realize the value they created, because it is the only way they can survive. Therefore, such consumer-generated information becomes a significant source of opportunities for firms to reinforce their value creation processes and to increase their survival likelihood. Recently increased attention to, so-called, Big Data, well reflects firms` innate desire for such information.
In this dissertation, we focus our attention on consumer-generated information analysis and applications development, particularly in the creative industries. First and foremost, the creative industries have become increasingly important to innovation and economic well-being. Almost all industries have tried to leverage their value by embedding creativity factors on existing technology and platform. Second, the creative industries is largely influenced by the IT invasion. The products of the creative industries have been easily converted and distributed in digital form. Finally, creative or cultural goods are by nature experiential, and thus consumers are usually not able to determine goods` quality ex ante. This is a significant feature of the industries in terms of applicability and usefulness of consumer-generated information analysis.
This dissertation consists of three separate essays. Each essay investigates how and what type of consumer-generated information can help producers to make better decisions in creative industries. First, Essay I analyzes the heterogeneous effect of word-of-mouth communication among consumers on box office revenue for mass and niche movies. Previous literature on WOM has consistent findings on the positive and significant effect of WOM volume on product sales, but the literature on WOM valence has been mixed. In this essay, we aim to explain the reason for the mixed effect of WOM valence on product sales by considering heterogeneous characteristics of products, especially in the movie market, by segmenting products into mainstream and non-mainstream movies. This study uses empirical data from the motion picture industry, such as box office revenue, WOM volume and valence, and other variables of movie characteristics. The hypothesis is tested using OLS and panel data analysis in econometric methods. We find a significant effect of WOM valence on box office revenue only in the case of non-mainstream movies, which have relatively smaller marketing budgets than mainstream movies. The findings suggest that as marketing communication channels become more diverse, with larger marketing budgets, the effect of online WOM valence on product sales can be diluted. In addition, it is found that the effect of WOM volume on box office revenue is greater for mainstream movies, suggesting that consumers build higher credibility on products with larger sales or WOM volume, especially for experience goods with uncertain quality. The findings explain the weak relationship between WOM valence and product sales, which has been controversial in the WOM literature, and broaden the understanding of the effect of WOM on product sales. The relationship between WOM valence and sales and, consequently, the revenue of a good has not been clearly understood, considering the heterogeneous characteristics of consumers in previous literature. In this study, it is found that WOM volume and valence have different effects on product sales, corresponding to differences in product category. The findings suggest a reason for the weak relationship between WOM valence and product sales, which has been controversial in the WOM literature.
Second, in Essay II, we propose a new technique of measuring user similarity in collaborative filtering using electric circuit analysis. Electric circuit analysis is used to measure the potential differences between nodes on an electric circuit. In this paper, by applying this method to transaction networks comprising users and items, i.e., user-item matrix, and by using the full information about the relationship structure of users in the perspective of item adoption, we overcome the limitations of one-to-one similarity calculation approach, such as the Pearson correlation, Tanimoto coefficient, and Hamming distance, in collaborative filtering. We found that electric circuit analysis can be successfully incorporated into recommender systems and has the potential to significantly enhance predictability, especially when combined with user-based collaborative filtering. We also propose four types of hybrid algorithms that combine the Pearson correlation method and electric circuit analysis. One of the algorithms exceeds the performance of the traditional collaborative filtering by 37.5\% at most. We hope that this work opens new opportunities for interdisciplinary research between physics and computer science and the development of new recommendation systems.
Finally, Essay III exhibits a customer-driven approach to analyze market structure of the digital music industry. Marketing practitioners often find it difficult to determine customers` preferences and identify the market structure of their products or brands. This essay presents a social media-based approach for market structure analysis in the digital music industry. We have applied a recently developed text mining technique to user-generated content and news articles to infer a perceptual network. Then, we empirically evaluated the impact of market structure variables on a musician`s chart performance by incorporating the perceptual network into a spatial model. The results robustly suggest that the method can successfully represent the market structure of musicians. In addition, we have developed a novel prediction method to better forecast the chart performance of musicians.