v-Fluence Blog

07/02/2006

There have been a variety of academic and commercial approaches to Web page categorization. Any framework automatically limits the scope of analysis but it is a necessity if one hopes to extrapolate a meaningful trend. As with any form of content analysis, the crucial question when judging the overall merit of a categorization scheme is whether the frame matches the research focus. In addition to this key point, Attardi et al. (1999) provide a list of criteria for judging categorization schemes: The list critiques a scheme on the basis of its overall quality, how concise and intuitive the scheme is; authoritativeness, how trustworthy the scheme is; accuracy, how appropriate the assignment of a document to a category is; consistency, whether similar documents are classified in similar ways; and completeness, whether all documents related to a given category are listed appropriately. These criteria provide a basic framework for analyzing some mainstream approaches towards Web page categorization.

In Relationship Marketing on the Internet, Roger Parker (2000) offers a scheme that categorizes Web sites in four ways. Inner-directed Web sites are primarily focused on accentuating a company’s brand through portrayal of client testimonials and company accomplishments; information-oriented Web sites are tailored towards providing the user with information about the company’s products and services; transaction-oriented Web sites are aimed at providing the user with the ability to purchase a commodity; and relationship-oriented Web sites attempt to engage the user by offering him/her opportunities to interact with the company through the site.

Parker’s scheme lacks the overall quality defined by Attardi et al., as Parker is excessively focused on Web sites as opposed to Web pages. This assumption neglects the fact that Web sites are multi-faceted. Most Web sites could be classified according to multiple categories in this scheme; and many company Web sites could be classified according to all four categories. One look at the v-Fluence.com Web site, for example, reveals a testimonials page that could be considered “inner-directed”; numerous pages dedicated to various products and services, which could be considered information-oriented; and a blog that provides the user with the ability to post feedback, which could be considered “relationship-oriented.”

Shifting the focus from Web sites to Web pages helps the scheme somewhat, but a closer examination reveals some inconsistencies and challenges for accurate categorization. For instance, inner-directed Web sites are typically “built from the business’ point of view,” whereas relationship Web sites “attempt to forge long-term bonds with Web site visitors.” Here, Parker categorizes one type of Web site based on its focus and another based on the goal it is attempting to accomplish. The scheme, therefore does not weight the same criteria consistently throughout each categorization and this creates challenges for anyone seeking to replicate the scheme.

Ultimately Parker is building a case for why “relationship-oriented” Web sites are the model that businesses should follow, and the other categorizations are defined in such a manner as to reflect their relative deficiencies. The value of this scheme appears predominantly as a marketing tool for Parker’s ideas on “relationship-oriented Web sites,” and its focus is primarily commercial.

The Drake Memorial Library at SUNY Brockport contains an alternative categorization scheme for Web page type. SUNY Brockport’s scheme differentiates between advocacy Web pages, which are administered by special-interest groups; business/marketing Web pages, which convey information about a company and the company’s products and services (similar to Parker’s categorization of information-oriented Web sites); informational Web pages, which usually belong to educational institutions or government organizations; news Web pages, which are hosted by organizations that provide news coverage; and finally, personal Web pages, which belong to individuals interested in promoting themselves and their interests.

This classification scheme is heavily focused on the entity or stakeholder involved with the administration of the Web site. That premise renders the scheme incomplete in its overall quality and authoritativeness. Although it may be concluded that a link exists between the administrator of the Web site and the site’s general intent, this link should not be used as the sole differentiating variable in the categorization of a Web page. What if a document prepared by a government organization is posted on the Web site of a non-governmental organization? According to this classification scheme, the document could be considered an advocacy Web page. Additionally, the scheme fails in terms of its completeness. For instance, how would a page with Bulletin Board System (BBS) be considered in this scheme? In addition, since the individual categories are defined by the agendas of the Web site administrator, the scheme neglects community-oriented Web sites and Web pages.

According to both of these schemes, a Web page can be compartmentalized into one category. This assumption produces substantial quality issues. In reality, Web pages are often multi-dimensional entities that can be categorized in various ways. A complete categorization scheme should encompass a number of different variables in defining Web page type. Ultimately, a categorization scheme should be tailored to meet the requirements of the basic research question and audience using the resulting information.

The following model can serve as an example of the process through which a categorization scheme for Web site type can be tailored to meet a very broad research question regarding what model of the Internet is prevalent within a sample population.