Category Archives: search

information retrieval, desktop search, personalized search, text analysis, search engines

GoogleBot vs iTunes Preview

Maybe Michael Bay should direct this next battle of the robot titans: Googlebot vs the Apple iTunes Web Servers – Dark of the Web?

It seems that as of today, the Apple iTunes Preview Web servers are not playing well with the googlebot. Take a look at the top three SERP descriptions for Angry Birds, one of the most popular iOS apps.

Google Search Engine Results Page for Angry Birds with useless description metadata

Maybe the Apple Webmasters need to get a plucky action hero to improve snippets with a meta description makeover.

"What is" Instant Search with Google and Bing

It seems there are some rather large differences in what I get when Bing and Google do their instant search term suggestions for something as vague as “what is”:

what is bing instant search suggestions what is google instant search suggestions

These results may also imply something about how much data has been gathered and is used for my personalized versions of both searches too. However, I don’t remember searching for any of these as either a "what is" search or for any of the other search terms.

Parliament? Gout? Gluten? The Illuminati? Strange indeed, but perhaps the makings of a great mystery-thriller novel!

Personalized Search

Personalized Search: A Contextual Computing Approach May Prove a Breakthrough in Personalized Search Efficiency


James Pitkow, Hinrich Schuetze, Todd A. Cass, Rob Cooley, Don Turnbull, Andy Edmonds, Eytan Adar, et al.


A contextual computing approach may prove a breakthrough in personalized search efficiency.


Contextual computing refers to the enhancement of a user’s interactions by understanding the user, the context, and the applications and information being used, typically across a wide set of user goals. Contextual computing is not just about modeling user preferences and behavior or embedding computation everywhere, it’s about actively adapting the computational environment – for each and every user – at each point of computation. (p 50)

The Outride system was designed to be a generalized architecture for the personalization of search across a variety of information ecologies.(p 52)

Search Engine - Average Task Completion Time in Seconds

While the results may seem overwhelmingly in favor of Outride, there are some issues to interpret. First, some of the scenarios contained tasks directly supported by the functionality provided by the Outride system, creating an advantage against the other search engines. Indeed, Outride features are specifically designed to understand users, provide support by the conceptual model and tasks users employ to search the Web, and to contextualize the application of search. This is the goal of contextual computing and why personalizing search makes sense.

Second, while the use of default profiles could have provided an advantage for Outride, it also could have negatively influenced the outcome, as the profile did not represent the test participants’ actual surfing pat- terns, nor were the participants intimately familiar with the content of the profiles. Third, some of the gains are likely due to the user interface since the Outride sidebar remains visible to users across all interac- tions, helping to preserve context and provide quick access to core search features. For example, while search engines require users to navigate back and forth between the list of search results and specific Web pages, Outride preserves context by keeping the search results open in the sidebar of the Web browser, making the contents of each search result accessible to the user with a single click. Still, the magnitude of the difference between the Outride system and the other engines is compelling, especially given that most search engines are less than 10% better than one another. (p 54)


information retrieval, search, information seeking, relevance feedback, personalization, contextual computing, user interfaces, search process

Cite As

Pitkow, J., Schutze, H., Cass, T., Cooley, R., Turnbull, D., Edmonds, A., et al. (2002). Personalized Search: A Contextual Computing Approach May Prove a Breakthrough in Personalized Search Efficiency. Communications of the ACM, 45(9), 50-55.

References in this publication

  • Anderson, J.R. Cognitive Psychology and Its Implications. Freeman, San Francisco, CA, 1980.
  • eTesting Labs. Google Web Search Engine Evaluation;
  • Pirolli, P. and Card, S.K. Psychological Review 106, 4 (1999), 643–675.
  • Gerard Salton , Michael J. McGill, Introduction to Modern Information Retrieval, McGraw-Hill, Inc., New York, NY, 1986

Publications that cite this publication

Advertising & Awareness with Sponsored Search: an exploratory study examining the effectiveness of Google AdWords at the local and global level

I will be giving a research talk (added recently, thus not on the conference Web page yet) titled: Advertising & Awareness with Sponsored Search:  an exploratory study examining the effectiveness of Google AdWords at the local and global level on October 28 at the American Society of Information Science & Technology (ASIS&T) 2008 Annual Meeting (AM08) in Columbus, Ohio.

This is the abstract for the talk:

This talk reviews an exploratory study of sponsored search advertising for a major US university’s academic department. The ad campaign used Google’s AdWord service with the goal of increasing awareness – not eCommerce – as part of the search process.  A behavioral model of information seeking is suggested that could be applied for selecting appropriate types of online advertising for awareness and other advertising goals. Insights into the study methodology will also be discussed including the use of increased integration with server logs, targeted site query terms and alternative awareness strategies. 

The talk is part of the panel AM08 2008 – The Google Online Marketing Challenge: A Multi-disciplinary Global Teaching and Learning Initiative Using Sponsored Search with Bernard Jansen, Mark A. Rosso, Dan Russell, Brian Detlor and Don Turnbull.

This is a summary of the panel:

Sponsored search is an innovative information searching paradigm. This panel will discuss a vehicle to explore this unique medium as an educational opportunity for students and professors. From February to May 2008, Google will run its first ever student competition in sponsored search, The Google Online Marketing Challenge Similar to other Google initiatives, the extent seems huge. Based on pre-registrations, more than two hundred professors and nearly nine thousand students from approximately 50 countries will compete. This may be the largest, worldwide educational course ever done. It is certainly on a large scale.

The Google Online Marketing Challenge is a real-life, problem-based, and multidisciplinary educational endeavor of the kind that many educators say is needed to relate teaching to outside the classroom. However, such endeavors are not without risks. The session should appeal to professors that competed in the 2008 Challenge, any professors considering the 2009 Challenge, as well as other educators who might consider the inclusion of Google AdWords as a pedagogical tool in their curricula. The panel will also be of great interest to those information professionals and educators as a possible model for use in other domains besides sponsored search.

Rating, Voting & Ranking: Designing for Collaboration & Consensus at CHI 2007

I’m in San Jose, California presenting a Works-in-Progress paper at the Association for Computing Machinery’s (ACM) Computer-Human Interface (CHI) 2007 conference. I’m showing off some of the interface design issues related to encouraging valid, fluid participation for a community-based internet content filter we’re developing at the University of Texas at Austin called OpenChoice.

Here’s the abstract for the paper:

The OpenChoice system, currently in development, is an open source, open access community rating and filtering service that would improve upon the utility of currently available Web content filters. The goal of OpenChoice is to encourage community involvement in making filtering classification more accurate and to increase awareness in the current approaches to content filtering. The design challenge for OpenChoice is to find the best interfaces for encouraging easy participation amongst a community of users, be it for voting, rating or discussing Web page content. This work in progress reviews some initial designs while reviewing best practices and designs from popular Web portals and community sites.

I’m also making it available to download: Turnbull, Don (2007) Rating, Voting & Ranking: Designing for Collaboration & Consensus. Works-in-Progress Paper presented at the ACM SIGCHI Conference. San Jose, CA. May 2, 2007.

Tagging 2.0 panel at SXSW2006 now a podcast

The Tagging 2.0 panel I organized at South by SouthWest 2006 in March is now a Tagging 2.0 podcast among the many SXSW 2006 podcasts you can download.

Some highlight quotes from the panel you really shouldn’t miss:

How can you pass up quips like that?

The Tagging 2.0 panel was one of the “highly-rated panels” this year, tied for first place with a number of other entertaining and informative panels, so check out their podcasts as they become available as well.