A couple of references on metadata and search engines

[Terry Kuny]

Hello everyone,

Sorry for the longish note on this *again* but
since the subject keeps coming up and the same advice
seems to be bandied about, perhaps these two recent 
will stir the pot a bit! ;-) 

-tk
---------------------------------------------------------

1. Heather Tunender and Jane Ervin. "How to Succeed in 
Promoting Your Web Site: The Impact of Search Engine 
Registration on Retrieval of a World Wide Web Site" 
Information Technology and Libraries, September 1998. 

- *none* of the engines indexed the META descriptor tag 
(Infoseek, AltaVista, Lycos, Yahoo!, Excite), even if 
they said they did.

- there are limits as to how deep spiders go 
(only Infoseek and Excite went below one level)

- some search engines never seemed to pick up 
revisions.

- YAHOO human indexers do not use or consider the 
META tag in their assigning of categories.

The study above ran for 4 weeks and the results for pickup 
of key phrases was very poor. None of the search engines 
appeared to index a planted keyword in META after 4 weeks.


TK: as an anecdotal aside, I have numerous pages on some large
internationally known resources which I have used META tags 
on for a couple of years, and have often checked to see 
whether 1) they were picked up and used in any meaningful way, 
and 2) if they made any difference in enhancing the retrieval
of said pages. The short answer is: no they are not and no
they don't.


2. User Interface Engineering did web-site studies in 
1998 and discovered that users looking for information 
on websites were less effective when they used search 
engines than if they had just followed links. Users found 
correct answers in 46% of tests BUT when they used the 
site search engine, the success rate dropped to 30%. 
When they used only links, they succeeded 53% of the time: 
Searching reduces success!

Garbage in/garbage out: classic rule applies. Most users 
did not understand how search engines might distinguish 
between partial and whole word searches. Most users did 
not understand the differences between plural and singular 
words in their searchers. Few searchers actually read the 
search instructions EVEN if they were directly below the 
keyword search box. Misspelled words are a big problem 
and people don’t always know they have misspelled a word. 

Problems interpreting search results were common. Most users 
had trouble determining why a search returned particular items. 
Most search results don’t provide any relevancy clues (like keyword 
in context). Some indexers use title fields and the titles 
of documents were poorly described.

Full-text searching is "a blunt instrument for chipping away 
at a large block of information." Indexes are more precise 
tools and good indexing is a skill. Humans are still much 
better at doing this than machines. Expect the development 
of good indexing services to become more common in the future.

"Searching is a difficult problem with no solution 
visible on the horizon." Their recommendation: "Until the technology is 
equal to the challenge, we suggest that designers seriously 
consider not including a search engine on their sites 
until the technology is equal to the challenge."

URL: world.std.com/~uieweb/searchart.htm

TK: Obviously findings like this about the success of local 
search engines could be extrapolated into the larger realm
of Internet search engines. 


---------------------------------------------------------------
Mr. Terry Kuny                  Phone:  819-776-6602   
XIST Inc. /                     Email:  terry.kuny at xist.com
Global Village Research         URL:    http://xist.com/kuny/

    Snail: Box 1141, St. B, Hull, Quebec, Canada J8X 3Y1
---------------------------------------------------------------