accessing repositories: Google versus the custom front-end

The November 2006 issue of D-Lib Magazine was released today. One point from an article on repository assessment caught my eye. For an IR implementation at the University of Wollongong [Australia], over 95 percent "of the measurable full text downloads of repository content" were accomplished through Google.

Having worked in the past year with enhancing repository access through Google for a regional project hosted at WSU, the data didn't take me completely by surprise.

I'm going to pick this up on Thursday by performing a similar analysis using the local repository data. In past months, I've built assessment reports describing repository access both through the custom interface and through search engines. I was always a little wary of the search engine data (thinking the counts must be too high), but Michael Organ's article is leading me to reconsider....