Dogpile comes out at the top of the pile

InformationWeek ran an article on October 30th titled “Survey Favors Dogpile’s Metasearch Approach.” The article reports that a recent study by J. D. Power and Associates, the customer satisfaction survey people, reports that Dogpile has ranked highest in customer satisfaction among search engines.

Dogpile is one of a number of popular metasearch engines. A metasearch engine is a kind of federated search engine that specializes in aggregating content from a number of search engines. (Note that noone refers to a search engine aggregator as a federated search engine but lots of people use the term “metasearch engine” to mean the same thing as “federated search engine.”)

Dogpile aggregates content from Google, Yahoo Search, Live Search, Ask.com, About, and other sources. According to the Directory and Search Engine History Dogpile came online in 1996, just a couple of years after the early search engines WebCrawler, Yahoo! and Lycos came into existence.

The J. D. Power and Associates Search Engine Ratings page shows Dogpile scoring 5 out of 5 for categories: Overall Satisfaction, Functionality, Ease of Use, and Results. Google Search, in second place, scored 4 out of 5 for each of the rating categories. The associated J. D. Powers Press Release announces that Dogpile has won the award for customer satisfaction among search engines two years in a row and that:

“Dogpile improves by 14 points since 2006 to earn a score of 818 on a 1,000-point scale. Dogpile performs particularly well among Internet service subscribers in all three factors that determine overall satisfaction (listed in order of importance): functionality, ease of use and results. Google follows Dogpile in the rankings with a score of 794, while Ask.com follows Google, earning a score of 784.”

Of course, the study isn’t saying that more people use Dogpile than Google. What it is saying is that Dogpile users are more satisfied with Dogpile than Google users are with Google. Note that the difference in points earned by Dogpile vs. Google is not huge: 818 vs. 794.

At first I didn’t find this study to be very interesting since my perception has always been that Google finds everything so why would I use Dogpile, or any other Metasearch engine, even if they were more “satisfying?” Well, an April 2007 study by Dogpile, in conjunction with Researchers from Queensland University of Technology and the Pennsylvania State University claims the opposite. The study looked for overlap of search results among the four most popular search engines: Google, Yahoo!, Windows Live, and Ask. The paper reports:

“The percent of total results unique to one search engine was established to be 88.3%.
The percent of total results shared by any two search engines was established to be 8.9% .
The percent of total results shared by three search engines was established to be 2.2%.
The percent of total results shared by the top four search engines was established to be 0.6%.
Interesting. Maybe I need to revisit the value of Metasearch engines. I still have questions, though. How does Dogpile perform relevance ranking among sources? How does Dogpile dedupe? If it does a really good job of both then it’s worth taking another looking at.