|
Tuesday, March 07, 2006
Prof. Jean Veronis of the Universite de Provence is writing on an important topic: perceived relevance of search results as delivered by major engines. The main finding is that Google and Yahoo are about tied -- with Yahoo going into the lead when Prof. Veronis follows Jakob Nielsen's suggestion to look more closely at how users actually scan pages. But the other conclusion is that all the engines score remarkably low on perceived relevance. Further, perceived relevance is lower with "commercial" links.
All of this makes intuitive sense. It's for reasons like this that Google has been trying to weed out "commercial" links on non-commercial queries, ever since Florida and in waves ever since. It's also the reason Google has been working on ad quality initiatives that seem draconian to some advertisers, but are intended to address the perceptions of users.
The primary problem with the study is the small number of evaluators. A group of only 14 postsecondary math and science students from France doesn't represent the French population very well. And in my recent post about global values and how these might relate to consumerism and markets, it's pretty clear that French values might be outliers in terms of philosophical outlooks on knowledge and science.
In that respect, then, this study is even less useful than the Stanford web credibility studies of several years ago, that for some reason used 50% Scandinavian participants, and 50% U.S.-based participants, including a high proportion of advanced undergraduates or graduate students.
Let's hope Prof. Veronis secures funding to do a proper study that includes a diverse set of evaluators.
Posted by
Andrew Goodman
View Posts by Category |
|
|
