Compete.com recently coined the term "search fulfillment" for this
metric
(http://blog.compete.com/2007/09/26/search-queries-results-yahoo-google-msn-live\
/).
They attempted to draw conclusions by comparing it across engines,
which is problematic, but internal to a site it's likely a good one.

I worked on generating this and similar metrics for MSN search until
Feb this year, and while I can't disclose raw numbers, this metric is
surpisingly low across the board.

Beware of the frequency of query refinements however, in the recent
press day at MSFT, bloggers report they stated that 50% of queries are
refined. A no-click, followed by a refined query + click is not quite
as bad as the no-click case.

A more more sensitive metric, from the Information Retrieval world,
the metric of Mean Reciprocal Rank is a good one -- the sum of
1/position clicked. This diagnoses ranking effectiveness.

Lou Rosenfield is due to release a book on search analytics, not sure
how focused he'll get on metrics, but looking forward to it:
http://www.rosenfeldmedia.com/books/searchanalytics/

Another potential confound, and one that plagues the Compete
conclusion, is the high frequency of navigational queries on major
search engines. If this is a case for your clients, beware of the
potential distortion caused by habitual query - click patterns for
navigation. On major engines, navigation queries may be as high as
30%: http://surfmind.com/muzings/?p=119



--- In webanalytics@yahoogroups.com, "Frank Reese" <reese.frank@...>
wrote:
>
> Hi,
>
> currently I'm working for some scientific search portals, websites for
> physicists, for ethnologists and so on.
> They nearly all have a thematic catalog and search fields to look for
> keywords in journals, books...
>
> To get an idea of how good the sites are working, I calculated for
example
> the ratio of 'visits with views of pages with search
> results' related to all visits and 'visits with detailed information
about
> information sources after a search' related to all visits.
>
> The results are not very encouraging. Does someone here has or knows
about
> information regarding the performance of similar websites?
>
> Kind Regards,
> Frank
>
> ---
> guide to web analytics solutions:
> www.idealobserver.com/solutionsguide
>
>
> [Non-text portions of this message have been removed]
>