Mining comparative opinions from customer reviews for Competitive Intelligence

Abstract

Competitive Intelligence is one of the key factors for enterprise risk management and decision support. However, the functions of Competitive Intelligence are often greatly restricted by the lack of sufficient information sources about the competitors. With the emergence of Web 2.0, the large numbers of customer-generated product reviews often contain information about competitors and have become a new source of mining Competitive Intelligence. In this study, we proposed a novel graphical model to extract and visualize comparative relations between products from customer reviews, with the interdependencies among relations taken into consideration, to help enterprises discover potential risks and further design new products and marketing strategies. Our experiments on a corpus of Amazon customer reviews show that our proposed method can extract comparative relations more accurately than the benchmark methods. Furthermore, this study opens a door to analyzing the rich consumer-generated data for enterprise risk management.

Introduction

Competitive Intelligence (CI) involves the early identification of potential risks and opportunities by gathering and analyzing information about the environment to support managers in making strategic decisions for an enterprise [33]. Most firms realize the importance of CI in enterprise risk management and decision support, and invest a large amount of money in CI. A survey from the American Futures Group consulting firm indicates that 82% of large enterprises and over 90% of the Forbes top 500 global firms adopt CI for risk management and decisions. By the end of the 20th century, the overall production value of CI industry had reached 70 billion U.S. dollars [23].

In order to identify potential risks, it is important for companies to collect and analyze information about their competitors' products and plans. Based on such information, a company can learn the relative weaknesses and strengths of its own products, and can then design new pointed products and campaigns to countervail those of its competitors. Traditionally, information about competitors has mainly come from press releases, such as analyst reports and trade journals, and recently also from competitors' websites and news sites. Unfortunately, such information is mostly generated by the company that produces the product. Therefore, the amount of available information is limited and its objectivity is questionable. The lack of sufficient and reliable information sources about competitors greatly restricts the capability of CI.

With the emergence of Web 2.0, an increasing number of customers now have opportunities to directly express their opinions and sentiments regarding products through various channels, such as online shopping sites, blogs, social network sites, forums, and so forth. These opinion data, coming directly from customers, become a natural information source for CI. There are some existing studies on mining customer opinions [6], [7], [27], [31], [34]. However, these studies mainly focus on identifying customers' sentiment polarities toward products. The most important problem in CI—i.e., collecting and analyzing the competitors' information to identify potential risks as early as possible and plan appropriate strategies—has not been well studied.

Customer reviews are often a rich source of comparison opinions. Users usually prefer to compare several competitive products with similar functions, for example,

• Nokia N95 has a stronger signal than iPhone.

• The iPhone has better looks, but a much higher price than the BB Curve.

• Compared with the v3, this V8 has a bigger body, and it has a much worse keyboard than Nokia E71.

These comparison opinions are precious information sources for identifying the relative strengths and weaknesses of products, analyzing the enterprise risk and threats from competitors, and further designing new products and business strategies.

Mining such comparison opinions is a non-trivial task due to the large amount of customer reviews and their informal style. In this paper, we propose a novel approach to extracting product comparative relations from customer reviews, and display the results as comparative relation maps for decision support in enterprise risk management.

The remainder of this paper is organized as follows: Section 2 reviews the related work in comparative opinion mining. Section 3 introduces our overall approach of comparative relation extraction. Section 4 introduces a novel graphical model we propose for comparative relation extraction. Section 5 presents our experiments that evaluate the proposed relation extraction approach. Section 6 concludes our study and discusses some future directions for research.