[Sigia-l] Outside In

[Lyle Kantrovich]

Interesting observations Paula.

A couple thoughts:

1. While optimizing search can be costly, we've learned a lot about
where to spend first in that area.  (See
http://crocolyle.blogspot.com/2004/04/search-best-bets-reporting-on-search.html
for just a couple references) Relying heavily on external engines
might severly limit a site owner's ability to collect useful data in
order to improve the site (content, navigation, AND search).

2. Of course site designers usually can't prohibit the use of external
engines - whether by employees or external users.  Because of this,
I've gotten into the habit of having usability test participants start
at *their* normal start/home page and then take whatever path is
natural to them.  Even if the site I'm testing isn't available via
their chosen engine/path, it's still very useful to learn how they
would naturally proceed -- where they'd go and what they'd search for.
 (If their path wouldn't work we eventually get them onto the site we
want to test of course.)

Testing this kind of "whole experience" is much like IBM's "out of the
box" testing (evaluating a new PC setup from the time it's delivered,
including taking it out of the box).  If we only look at the
experience from the time the user gets to the home page we can miss
some really key points...like that people often land somewhere other
than the home page...or that they'd never even think that something
like your site / company / product exists (yikes!).

On the research side, again see some of the links I blogged at
http://crocolyle.blogspot.com/2004/04/search-best-bets-reporting-on-search.html
 -- especially Lou's presentation on Search Log Analysis.  The zipf
distribution point was *huge* for me.  No longer do we have to tag all
kinds of content...rather monitor your top search terms and optimize
results for them (first) using things like synonyms and best bets.

--
Lyle

--------------------------
Lyle Kantrovich
Blog: Croc O' Lyle
http://crocolyle.blogspot.com