[Web4lib] Would an open source Best Bets tool be useful for libraries?

[Richard Wiggins]

Lou Rosenfeld and I are working on a book on search analytics.  The first
chapter of the book describes how I fell into search analytics and the Best
Bets / search analytics paradigm.  With Best Bets, if someone searches your
home page for "library hours" or "special collections" or "newspaper
database" or "JSTOR" you guarantee that the search gives "the" best hit at
the top of the hit list, because you manually edited the entry into your
database.

You choose the entries you put in the Best Bets database by analyzing your
search logs.  You take the top 1000 or so popular searches and you hand-code
links from each search phrase to the best URLs.

See
http://www.rosenfeldmedia.com/books/searchanalytics/content/sample_chapter
for a sample chapter about analytics and Best Bets.

I'm posting to ask if a free Best Bets tool would be of use to libraries in
general, and research/university libraries in particular.  There are lots of
alternatives.  Some libraries might make their catalog the default search,
and insert the Best Bet URLs into the catalog.  Others might use Google
Appliance. or Ultraseek, or the new Yahoo / IBM Omnifind free search engine
(up to 500,000 pages), all of which have Best Bets features.  Others might
use faceted or federated search tools they've implemented.

When we built our Best Bets tool at Michigan State, we chose ASP and
MS-SQL.  For a couple of years I've dreamed of re-implementing as an open
source LAMP (Linux/Apache/MySQL/PHP) tool.  It would include a search
logging function.

So: might a free Best Bets tool be useful to you?

Thanks,

/rich

PS -- If you will be at the IA Summit in Las Vegas next week and have ideas
on this question, please chat with one of us.