[Sigia-l] ROI/Value of Search Engine Design - Resources?

[Richard Wiggins]

Once I saw a cartoon that showed a fellow with a bow and arrow.  He'd shot
several times into the side of a barn.  Every shot was a bull's eye.  He
also had a bucket of paint and a paint brush.  "It's easy," he explained. 
"First you shoot the arrow, and then you paint the target."

Now that Jared has exposed his ultimate biases: <quote>...I believe that
Search is essentially an experiment that has failed, and, along with voice
input, Eliza-style conversation bots, video conferencing, and little
animated characters who offer you unwanted help when you're trying to type a
letter, designers will only use Search in highly-specialized, fringe
applications and the occasional retro art piece.</quote>

... it is not at all surprising that his experiments never reveal situations
where Search (proper noun?) is useful or merits content provider care and
feeding.  

Let's posit for a moment that we are all headed for a future where we
navigate some whiz-bang multidimensional space instead of typing "Linksys
WiFi" or "rule against perpetuities" into a search box. (I'm picturing
shooting letters out, arcade-style, one at a time: L-I-N-K...) I don't
believe it for a minute, but let's just assume it. 

That technology isn't here yet. (If you think it is, quick, go sell it to
Yahoo, because their drill-down service loses market share to Google every
day.)  While we wait for that technology, we have to live in the here and
now and make the most of the tools we do have.

This discussion reminds me of the sudden push for hydrogen fuel cell cars.
It's a technology people have been talking about for a couple decades, and
some folks claim it's right around the corner -- as they've claimed for
years.  But fuel cell technology is unproven, and there are proven
technologies such as hybrids already in the streets.  So for my money you
make the most of the proven technologies even if you're confident the
Jetsons will be here in 20 years.

So if you think search engines someday will end up in the dustbin of history
with bloodletting leaches, the steam engine, and Eliza, fine.  I think the
proposition is ridiculous, but you're entitled to your visions. I'm still
looking forward to the useful version of the 3D OPAC I saw in 1994. Frankly,
both the AI camps and the LIS camps have not delivered on their many
visions.  

You can't let visions limit what you accomplish in the here-and-now.  In
2003, lots of sites offer search engines, and lots of customers use them. 
You can make better choices about search, or worse ones.  The better ones
serve the customers who choose to search better.  And features such as Best
Bets have more payoff, and less investment up front, than people realize.

Today, in the here and now, I continue to strongly believe:

-- Some customers leap to a search box if it's available.  Percents vary,
but some people prefer search over clicking (in today's world) so why not
serve them as well as you can?

-- Every Web site that's non trivial in size needs a search box on the home
page.  We can quibble about what non-trivial is, but certainly a site with
thousands of pages needs a search box on the home page.

-- If you want to honestly appraise whether customers benefit from Search,
you need to do a double-blind test. Take a site that has reasonably good
link structure, and present two versions of it, with and without a search
box on the home page.  Use a decent search engine, including Best Bets. 
Give two sets of guinea pigs the same list of tasks (in writing; no
coaching). Measure time to accomplish tasks and guinea pig satisfaction. 
Repeat for several types of sites.  

Saying "the clothing sites without search performed better in our tests"
doesn't make the case.  Maybe the ones without search knew they had to do a
lot more work on categories and labels.  Maybe they tend to be smaller. 
Maybe all clothes from the Gap devolve to the same item.  Unless you measure
the Gap site with and without search, you haven't controlled all the
variables.

Finally, Jared, since you're into categories: now that I know your biases,
please move me from the "perplexed" category to "bemused".  Until the
Jetsons are here, I'll take my usability of search cues from Jakob Nielsen.

Over and out,

/rich

On Wed, 19 Feb 2003 13:52:15 -0800 (PST), Karl Fast wrote:

> > You're absolutely correct -- I really do see the future in terms of
> > categories and clicking. The more I watch what's happening with the
> > evolution of web sites, the more I believe that Search is
> > essentially an experiment that has failed....<snip>
> 
> From my own work (which is more limited than yours) I have found
> myself leaning towards this view. I'm not as fervent as you, Jared,
> but I am definitely warming to your idea (my readings in information
> visualization has pushed me over the edge).
> 
> Adopting this view suggests something interesting:
> 
> 1. The LIS concept of controlled vocabularies is of critical
>    importance to making this all work (by CV I included facets,
>    descriptive cataloguing, authority control, blah blah blah).
> 
> 2. LIS has had "the answer" for a long time (maybe "the answer" is
>    too strong; how about "key element")
> 
> 3. LIS has failed to capitalize on this. One need look no further
>    than OPACs (electronic card catalogues) to see how badly they
>    have missed the mark.
> 
> In other words, LIS has figured out how to create a semantic layer
> and that this layer is the key, but they've done a lousy job of
> making that layer useful through an intuitive interaction layer.
> 
> 
> 
> --karl
> ------------
> When replying, please *trim your post* as much as possible.
> *Plain text, please; NO Attachments
> 
> ASIST IA 03 Summit: Making Connections
> <a
href="http://mail.richardwiggins.com/jump/http://www.asist-events.org/IASummit2003">http://www.asist-events.org/IASummit2003</a>/
> 
> Searchable list archive:   <a
href="http://mail.richardwiggins.com/jump/http://www.info-arch.org/lists/sigia-l">http://www.info-arch.org/lists/sigia-l</a>/
> ________________________________________
> Sigia-l mailing list -- post to: Sigia-l at asis.org
> Changes to subscription: <a
href="http://mail.richardwiggins.com/jump/http://mail.asis.org/mailman/listinfo/sigia-l">http://mail.asis.org/mailman/listinfo/sigia-l</a>

____________________________________________________
Richard Wiggins
Writing, Speaking, and Consulting on Internet Topics
rich at richardwiggins.com       www.richardwiggins.com