Applied Information Management
 
News and Updates

Using content tagging and recommender systems to enhance collecting relevant user data from social software

In Brief: The Internet is all about information—users looking for it and organizations looking to collect it. In fact, web sites battle daily for customers and users by using social software outfitted with sophisticated methods of collecting user-specific information so they can personalize content and online experiences.

Mining valuable user information online quickly and accurately is challenging, yet essential, for organizations and consumers.

But this ability to pinpoint the right information, called findability, meets with different levels of success depending on the method used to collect this information, or metadata. This study helps organizations understand two of the most popular methods so they can choose the one to best suit their needs.

The first relies on recommender systems that use a specific filtering technique that attempts to present choices, whether for movies, music, books, or news, based on user input. For example, user product ratings, rankings and favorites lists offer insight into preferences, which are then compared to similar data from other users before presenting recommendations to the user.

On the other hand, content tagging systems follow a user's own pattern rather than using preexisting formats. By collecting data, such as which photo, news story, or blog a user tagged for later or ongoing reference, it's possible to create an "interests" profile that organizations can use to tailor content.

The ultimate goal of social software and social computing, of course, is to actively involve users in creating, classifying, and discovering information. To that end, this research shows that content tagging systems excel at classifying data, while recommender systems prove strongest in discovering information.

Thanks to a detailed summary of each system's strengths and weaknesses, coupled with examples of applications that take advantage of these systems, this work is invaluable as a comparative tool for user experience practitioners who design web-based social software applications.

Content Tagging System and Information Classification
Strengths
  • User-classified content is personally and contextually relevant
  • Captures unique user and community generated vocabulary, specialized terminology, and jargon
  • Easy to use—low barrier to entry, low cognitive/learning costs
  • Incentive—users typically classify for their own personal needs while their tags benefit all system users.
Weaknesses
  • Uncontrolled, ambiguous, and unreliable categorization
  • Requires a large population of classifiers to generate useful results
  • Allows for classifier term ambiguity—spelling errors, multiple word tag descriptions, and personal notes such as "todo" "toread"

Figure 1—Strengths and weaknesses of content tagging system and information classification

References

  • Bielenberg, K & Zacher, M. (2005). Groups in social software: Utilizing tagging to integrate individual contexts for social navigation. Master's Thesis in Digital Media. University of Bremen, Germany.
  • Dye, J. (2006). Folksonomy: A game of high-tech (and high stakes) tag. EContent, 29(3), 38-43.
  • Golder, S. A., & Huberman, B. A. (2005). The structure of collaborative tagging systems. Information Dynamics Lab: HP Labs, Palo Alto, USA, available at: The Structure of Collaborative Tagging Systems
  • Herlocker, J., Konstan, J., Terveen, G., Riedl, J. (2004). Evaluating collaborative filtering recommender systems. ACM Transactions on Information Systems, (22) 1, 5-53.
  • Marlow, C., Naaman, M., boyd, d., & Davis, M. (2006). HT06, tagging paper, taxonomy, Flickr, academic article, toread. Proceedings of the seventeenth conference on Hypertext and Hypermedia, Odense, Denmark. New York: ACM Press, 2006.
  • Morville, P. (2005). Ambient findability: What we find changes who we become. Sebastopol: O'Reilly Media.
  • Resnick, P. & Varian, H. R. (1997). Recommender systems. Communications of the ACM, (40)3, 56-58.
AIM alumnus Michael Barnes

Research Paper Author: Michael Barnes—2007 AIM Graduate, Manager, User Experience Design, VeriSign, Inc.

Abstract: This study describes how user-generated metadata may be leveraged to enhance findability in web-based social software applications (Morville, 2005). Two interaction design systems, content tagging (Golder & Huberman, 2005) and recommender systems (Resnick & Varian, 1997), are examined to identify strengths and weaknesses along three findability factors: information classification, information retrieval, and information discovery. Greater overall findability strength may be found in content tagging systems than in recommender systems.

Download the entire Capstone research project