It is now perhaps a cliche to say that the nature of business has changed. With the rise of the Internet, increased globalization, and advances in financial engineering, both large and small companies have to keep abreast of trends at shorter and shorter time scales. The volatility of the marketplace has sent the business environment into a roller coaster ride of confidence and anxiety, and so it is imperative that sophisticated tools be brought in to handle gigantic market swings. These tools can take the form of software, and go by the name of competitive intelligence (CI). The goal of this book is to evaluate this software based on certain criteria that the authors think is important. It is one of the first books on the evaluation of competitive intelligence software, and is successful in cutting through the advertising hype that frequently accompanies this software.
Competitive intelligence for the authors is a process that involves using publicly available information in order to learn various things about competitor, and to understand this information thoroughly. They assert, correctly, that the information that is gained must be transformed in order to make it useful for decision making and to induce changes or actions in a particular company.
In chapter 1, the authors attempt to clarify the meaning of "value-added information" in the use of CI. One would think that this would be a difficult notion to clarify, and this is certainly correct. The authors approach the problem by attempting to define just what "information" and "intelligence" are, as well as "data" and "knowledge". Such definitions could be deep and might degenerate into philosophical discussion, but the authors do a fairly good job of keeping the discussion relatively concrete. This discussion leads them to distinguish between the roles played by information specialists, CI professionals, and experts. An information specialist acquires access to information resources, CI practitioners assign values to its content, and experts decide the action to be taken. The authors though recognize that the boundaries between these roles can be blurred. After a modest review of the literature, the authors assert that value-added processes are ones that offer the means to see the potential of information and to relate it to problems in specific environments.
The authors attempt to construct a conceptual framework for CI in chapter 2, after giving a literature survey of attempts to do so. As expected, the laissez faire nature of industry in the US made the nature of its CI very different from the CI of Japan or Europe. The authors are careful to distinguish between CI and "industrial spying", and clarify the difference between it and business and marketing intelligence. Different analytical techniques, such as personality profiling and scenario development, are discussed in terms of their ability to guide information requirements. In addition, they emphasize the need for reliable filtering mechanisms that will eliminate false information about a competitor. The most interesting discussion in this chapter concerns the analysis of the obtained information, for this is where techniques from artificial intelligence could be used. Such techniques are not discussed in the book, but the authors do summarize the eight most popular analysis techniques for CI.
In chapter 3, the authors begin their evaluation of CI software, with the main goal being to find out whether it can allow users to achieve their intended goals. Their evaluation criteria are aimed at identifying the value-added processes that should take place when a CI application is used to transform information into intelligence. The authors stress early on that CI software needs to be evaluated beyond the "recall" and "precision" criteria used to evaluate information retrieval systems. The dynamical nature of competitive information is the main reason for this, as typical databases are not refreshed at short enough time scales. CI systems also must assist in the analysis of information, not merely retrieve it. The value-added framework of R.S. Taylor, one of the early CI specialists, is used throughout this chapter, and the rest of the book, to evaluate CI software. Based on the Taylor model, the author presents 38 criteria for evaluating CI software, and discuss them in fair detail. One of these criteria is particularly interesting, in that it involves "closeness to the problem", a very difficult concept to quantify, but one which is also very important in other fields, such as artificial intelligence. And, by the way, the use of artificial intelligence will soften the need for a "sixth sense" that the authors mention is a necessary ability for CI specialists to have in order to analyze information. Indeed, recent advances in natural language "paraphrasing" will be of enormous importance in the need for summarizing acquired information.
Finally, in chapter 4, the authors begin evaluating the software packages available for CI. The authors list three selection criteria for distinguishing CI from other types of software, and six applications that meet these criteria. In chapter 5 the authors present a set of equations to allow more rigorous evaluation of CI software. Their goal was to compare these packages relative to their information-processing capability, and not rank them. It is readily apparent when reading this chapter that the authors took great care in their evaluation of the packages, which certainly must have been a time-consuming effort. Many problems shared by all the packages are discussed, including their lack of tools for monitoring the relevance of content through time, the lack of mechanisms for filtering information, and the poor performance of the packages when dealing with acquisition of knowledge. In addition, the authors conclude that the analytical capabilities of the packages, i.e. their ability to transform information into intelligence, are almost non-existent. Such capabilities, they argue, require human intelligence, and this is an interesting comment if comparison is made to recent advances in artificial intelligence. The authors remark that these packages are far from being intelligent, and that such intelligence is needed in order to make CI a viable technology, which in their opinion currently is not.