This article was written before the Search Engine Placement Index
was developed although still accurate as an overview when it comes to improving your
placement the "Search Engine Placements Secrets"
site is the best place to find this information.
Overview of Search Engine Placement
Not unlike a bad marriage where the couple is staying
together for the sake of the children is what I liken the current relationship between
webmasters and search engines to. Each is dependent upon the other to provide a service.
The search engines need the developers to inform and allow them to index their work and
the webmasters and site designers need the search engines to index their sites for the
exposure. Since this is a co-dependent situation for the life of me I can't figure out why
there is an adversarial relationship between the two.
The search engines exacerbate the situation by providing
little or no useful information about how the rankings are decided. This leads to
"keyword and submission spamming" a major no no with most engines. Since a good
search engine placements or positions is important to the developer for two reasons, those
being client satisfaction and exposure. Who can blame them!
However, the search engines also owe it to their patrons to
make sure that the rankings are a true rating of the content contained on a site. Anyone
who uses search engines extensively for research can relate to this statement as this is a
problem frequently encountered by them. The reason for unrealistic ratings or misleading
ratings are the practices mentioned in the preceding paragraph.
As an example I recently was interested in using Pegasus
mail software. Using "pegasus mail software" as a query at five of the search
engines resulted in thousands of matches and Pegasus did not even score in the top fifty
in four of the five. I did however find that the address is pegasus.usa.com from an
article at a university demonstrating usage and uses of the software. Now I will discuss
the results from five of the top engines and how I feel that the results were arrived at
for the top ranking.
Top rating was for Que E-mail. This site had no meta tags
whatsoever and the title was Que E-Mail Software. They did however have in the body of the
document a mention of Pegasus as it turns out that this site was enabling a download of
several free and shareware versions of mail programs. So, the end result was an ability to
download a copy of Pegasus but no actual contact with Pegasus for additional information
or the actual url.
In order to ascertain how the ratings are generated I went
to the help area for submissions. No information regarding ratings was available. They do
however mention that they come back and index your whole site at a later date. They also
added that they have two categories that they "manually" rate your site for
inclusion or to put it another way an actual person rates the site. These are Lycos A2Z
Sites By Subject and Lycos Point's Top 5% Sites.
Judging by the results from this query I would deduce that
rankings are generated primarily from the content of the page. I am also sure that the
title is used here as a recent e-mail from a listee of a list I subscribe to mentioned
that <!--Your title repeated like this-->, a title trick recommended in a few
articles I have read results in a penalty at Lycos and Alta Vista. Also keep in mind that
it is not just the title matching your query but the title and content contained on the
site matching the title on the site as well. Matching keywords to title and query is not
considered here as meta tags don't seem to be used by Lycos.
Top rating was for Pegasus Airwave Incorporated. This site
sells therapeutic products. There were meta tags, however, keywords meta tags were only
found on the index page there were only developer and author tags on the pages found
deeper in the site. The title used is Pegasus Airwave Incorporated. This site was even
The keywords were in alphabetical order so the placement or
spacing of the keywords is not paramount as suggested in some of my research articles. Of
most interest to me was that the index page contained only a logo and one of those
wonderful Looks Best logos!! The meta tag for keywords contained 259 words and only
matched the keywords one time each in the keywords tag. All title words however were
matched in the keywords, 2 matched in the developers tag and there were three instances
where the keywords were used in image tags. There were only 2 matches in the content that
is viewed by the user. I also noted that there is a meta refresh tag which in some
articles I read noted that some robots will not index them, doesn't seem to be the case
here. Also of note is that there were no comas between keywords. This is important because
the search engines and validators only allow 1024 characters, in a long list this would
enable you to add several more words!!
The help section of the submission area told briefly when
they will index you and how but no information about how rankings are generated is
available. I did learn from this that engine uses a type of fuzzy logic that allows some
words not necessarily part of your query to count in ratings. For instance a query for
poodles would include dogs even though not queried. This is the crux of this companies
high rating as two of their keywords were program and post which using their fuzzy logic
were used to replace software and mail giving it a 100 % match. The bottom line here is
that matches to title, and query are important content is not considered at least not on
this index page. This would also lead me to believe that a normal search at this engine
produces a lot of useless matches. Also the letter starting your title does not influence
rating. For instance by saying A...... would not get a higher rating.
Excite supplies full disclosure of how ratings
are arrived at. Their system is a match of the query to the content of the
page. In their information there are warnings against "submission and word
spamming". " Word spamming" doesn't work because their robot makes sure that
all content counted can be read or viewed by the user! They seem very
interested in supplying true ratings to their patrons. I would like to
note that an article that I read about getting good search results rated Excite at or near
the top. All said and done Pegasus did not place anywhere near the top in the ratings,
however, a page on installing and configuring Pegasus Mail did score at the top! I will
discuss what this means to the webmaster and the user a little later in the summary.
The top ranking here was Universe Software which offers
downloads and information on several mail clients of which Pegasus was one. The title of
the page is Universe Software-Mail Clients. Almost all the top ten getting top ranking at
Infoseek were selling or offering information about mail software. Of special interest at
Infoseek is their disclosure of how relevancy ratings are arrived at.
Use a highly descriptive title, include a META tag description, and create META tag
keywords that contain comma-separated phrases. Use an assortment of synonyms that
accurately describe your site, but don't try to boost your site's relevance by repeating
keywords. The overuse and repetition of keywords may result in a lower relevancy score and
possible omission from Infoseek's index.
First off I would like to bring to your attention this fact,
the fourth (third was a 404 error) site in the site relevancy rating had no meta tags
whatsoever but had the query words sprinkled liberally through the document including the
word Pegasus several times even though it was a page concerned with configuring Eudora.
Two of the query words were in the title. Also of note is that the number 1 rated site had
the word mail in about 70% of the keyword phrases with little spacing between them. So,
keyword spacing isn't important and "spamming" is not necessarily penalized. On
the second rated site all the queried words were contained in the meta tags however
nowhere in the actual content of the document was the queried word Pegasus to be found.
The FAQ that Infoseek provided is very misleading. Judging
by the rankings they are using a combination of meta tag and queried words contained in
the document to formulate the rankings with a slight emphasis on meta tags. Title does not
seem to be as important as the FAQ had implied. It is this type of behavior by Infoseek
that encourages tricks by webmasters to get a higher rating of relevancy then deserved.
The losers are the users of Infoseek and the webmasters that submit an honest document
containing keywords that are relevant and fair!
By far the favorite search engine (?) among internet users.
Yahoo actually started as an indexer of sites and did not add the search engine until
later. Part of their success is due to the accuracy or quality of results from a query and
partly due to one of the best promotion campaigns on and off of the net that I have ever
seen!! It will be a long time before this company ever generates enough income to warrant
the absurd prices that were paid for the stock when issued.
All of this combined with an extremely effective internet ad
campaign which has slowed somewhat lately has made them the Kings of Search on the Net!!
Almost all developers will tell you that in most cases Yahoo is the only search engine
that customers seem to care about. You could get top ten status in the four other top
engines mentioned here and if you were way down in Yahoo the client would most likely be
unhappy. The site developer would be scratching his head but a developer in the know would
understand. The reason being, Yahoo is responsible for around 45% OF ALL TRAFFIC TO MOST
SITES!! Well I digress but I felt that a little background on this was warranted.
The result from Yahoo resulted in Pegasus Software being the
highest rated in three of the five results that were found. This illustrates that the
indexing of sites into categories by far results in the most accurate results however if
you aren't sure of a category for a subject it makes this a far more arduous procedure.
To get a better idea of how ratings come about here I
decided to try a regional search in a category that would generate loads of matches. I
queried for internet presence providers in the GTA region of Ontario. The results were
that the number 1 ranking went to The Banner Shack. Judging by the facts gathered here ie:
no keyword meta tags only a description meta which didn't match the description on the
Yahoo site. I conclude that Yahoo uses the submission form almost exclusively to index a
site. Also keep in mind that Yahoo is one of the very few that actually manually checks a
pages content for relevancy to a topic or category. The placement is also entirely
alphabetized by company name with the title on the page being ignored by Yahoo. The bottom
line here is that there is not much of a chance to manipulate placement on Yahoo.
To sum up this little exercise the search engines that
supply the best and most accurate information to webmaster serve their patrons best by
producing accurate ratings of relevancy results. Those that play the adversarial role with
webmasters and site developers encourage them to play the games that are to blame for most
of the animosity between these two and ratings that are out in left field compared to the
actual content at a listed site. The search engines would be wise to follow the lead of
Excite and Yahoo and use full and truthful disclosure as to how ratings of relevancy are
arrived at as well as examine the methods they are presently using to index and list
sites. It seems like there is a large pile of manure to weed through before you get to the
real information you're looking for. A recent survey by Georgia Tech showed this to be a
common complaint of users.
I have been gathering information on this subject for over
three months and it was this information which spurred my desire to do this article and
little experiment. To follow up this article in the next two issues I will be doing some
research and assembling an overview of the twelve articles that I have found to date on
the net. I will also be doing a submission blitz to see what kind of placement I can
achieve for this site at the various search engines and directories using the information
gained from this exercise and the overview.