[WEB4LIB] Amazon vs. Google as our corporate role model (was: In defense of

[JOHN MARQUETTE]

Amazon's results are better most likely because the company wants the
user to spend money, preferably on an item closest to the user's
original intent, and capture that user so s/he becomes a regular
customer.

Google's ranked results are based on their own proprietary - and quite
clever - algorithms which not only show are designed to get a user to
click through to a search-related site.

In the academic and public library environments, we don't have any
expectation of profit since we're here to serve the public good or the
mandates of our trustees.  However, our users are often faced with
higher expectations (grades, which can vary according to the quality of
information users receive, for example) than shoppers (who look for
known items with fixed characteristics) or people looking for Britney
Spears (or digital manifestations of her in photographic, textual, and
audio formats).  

Absent the profit motive and with only the desire to provide goodwill
remaining to us in the LIS professional community, what will motivate us
and our organizations to provide uniform results for all users
regardless of access method?  Isn't that everybody's goal, after all?

John Marquette

-----Original Message-----
From: [mailto:web4lib at webjunction.org] On Behalf Of Peter Murray
Sent: Tuesday, May 10, 2005 5:26 PM
To: Multiple recipients of list
Subject: [WEB4LIB] Amazon vs. Google as our corporate role model (was:
In defense of

This was part of a discussion last week, but the Web4Lib mailing list 
was eating my reply because of the PGP signature.


Peter

- -------- Original Message --------
Date: Fri, 06 May 2005 14:52:38 -0400
From: Peter Murray <peter at OhioLINK.edu>
To: Multiple recipients of list <web4lib at webjunction.org>
Subject: Amazon vs. Google as our corporate mentor (was: In defense of
stupid users)

If we take as a given that users want to "find" more than they want to
"search", I'd argue that holding Google up as the commercial version of
what we strive to be falls short of the mark.  It can be argued that
Amazon is a more successful interface than Google.  For example, if you
were looking for something about "Jack and the Beanstalk", which of
these would you prefer?

Google
http://www.google.com/search?hl=en&q=Jack+and+the+Beanstalk&btnG=Google+
Search

Amazon
http://www.amazon.com/exec/obidos/external-search?field-keywords=Jack%20
and%20the%20Beanstalk&index=blended

Google's results is just a flat 1-260,000 "relevance ranked" list of
links.  Amazon's results, the functional equivalent of performing the
broadest possible search in Amazon's single search box, shows you the
most popular options plus allows you to "Refine Your Search" into areas
such as Industrial Supplies and Medical Supplies.  (Believe it or not,
at the time this message was crafted, there is an entry for one "Jack
and the Beanstalk" medical supply.)  So which "Simple Search" would you
rather go to if you were searching for Jack and the Beanstalk Medical
Supplies?

The funny thing is that we have a rich source of information already in
our databases.  Are tools like LCSH and Dewey really worth the effort it
takes to maintaining those systems and applying them to our items when
we don't take full advantage of them?

We're building a new object repository in Ohio (the Digital Resource
Commons -- if you are interested in helping build it, see the job
opportunity announcement sent out yesterday).  I'm striving to make that
repository's search box behave more like Amazon than Google.  We'll make
use of the system's knowledge of media types to offer the breakdown of
results into the equivalent of "Movies" and "Books".  We're also going
to take advantage of a repository-wide general taxonomy and its links to
community-specific taxonomies (and maybe even "folksonomies") to provide
the faceted results to the user.  Perhaps it can take advantage of a
public comments engine to affect the order of search results, as in a
recommendation system.  Or even track (anonymously, of course) past
users' navigation paths through the system and offer "Related Searches."

In short, the commerce world has figured out that the intelligence needs
to be build into the back end (e.g. the search engine), not into the
front end (e.g. instruction).  If we don't do that, we'll continue to
drive users to the interfaces and services that do.


Peter
- --
Peter Murray                       http://www.pandc.org/peter/work/
Assistant Director, Multimedia Systems  tel:+1-614-728-3600;ext=338
OhioLINK: the Ohio Library and Information Network   Columbus, Ohio