Sunday, March 8

The Truth About Search Engines - Part 2

WWW's "historical" logo, created by ...Image via Wikipedia

Robot Travel

If there is one thing I have learned about robots, it is that
there is absolutely no pattern to them. Most robots are stupid
and wander randomly. For example, 50% of robot hits to my
sites, ask for the robots.txt page and then go away never
asking for anything else. Then they come back a week later,
ask for the same thing and then go away, again. This happens
over and over again for months. I have never figured it out.
What are they doing? If they wanted to see if the website was
really a web site, they could just Ping it. This would be much
faster and much more efficient. They seldom visit another page
and if they do, they ask for one other page every visit or so.
Some come in and issue rapid-fire requests for every page in
the website. How rude! You have to quit worrying so much about
robots. It takes 6 months before they request enough pages to
do you any good. I really quit thinking about them a long time
ago. Build a lot of pages correctly and, if you have reciprocal
links to them, the robots will find them someday.
Try this: Go to AltaVista and type into the search box link:
YourSite.com (Leave off the www). This will list the reciprocal
links to your web site. Try link:crownjewels.com and you get
136 links to it. Think about this now: The robots say to
themselves, "Here is a site that must be popular or why
would so many websites SIMILAR to it have it's link on their
pages?" Remember that only SIMILAR sites with SIMILAR
THEMES would probably have a link to your site. They give more
importance to this than you submitting your link to them.
Wouldn't you?
Go to heavily trafficked sites matching your web site's Themes
and use AltaVista to find out how many reciprocal links they
have. This will prove to you I am right.
Search engines are nothing more than links to your site. The
problem is, you are constantly having to fight for your
positioning in the search query listings. Forget about that.
Leave the fighting to people who are able to spend 24 hours a
day trying to trick everybody. Quit trying to compete with the
large organizations pouring millions into their marketing.
Completely forget about Search Engine positioning and go after
the reciprocal links. The Search Engines will then believe you
are a heavily visited site because you will be. You will now
be getting the traffic you so richly deserve.
Search engine visitors to your site, are often-times not
qualified visitors. Too many visitors pop into your home page
for 2 seconds and then leave. You know how it is. We all do
it when we are using the search engines. Either it wasn't the
information we were looking for, or they had this huge graphic
on this stupid portal page, which just took forever to load.
These visitors shouldn't even count, but they get counted as
12-18 hits in your server logs. Hits are requests to the
server. One page request can incur a lot of hits: requests to
the page itself plus the graphics, each count as a hit.
Reciprocal links bring in qualified visitors. These are
visitors who were already on a web site which had matching
Themes to yours. They already have a good idea of what type
of site you are. They will come into your site and actually
stay awhile. These visitors should count as double credit,
they are so good.
I know which type of visitor I would rather have.
How do you get people to WANT to put your link on their web
sites? Why would a similar site put a link to your site on
theirs? Simple, you have similar Themes. You are similar,
but not competition.
There is one very important lesson to be learned from this
crazy robot behavior. You need to make the navigation in your
web site so easy that a visitor can find any page within 2
clicks of your home page. One way of doing this is installing
hidden DotLinks. Dotlinks are little periods that are linked
to other pages which are not really noticeable on your page
if you put it as a period. Although they are not easily seen
by the human eye, they are a link that a robot can follow] in
your web site. When you do this, robots can find your pages
faster and more easily.
Giving the Robots What They Want
So how do you make the search engine robots give your site
a better rating than all the other millions of websites trying
to do the same thing? Simple, give them what they want. You
can't trick them or make them think that you are better than
you are. Think about a visit from the eyes of a robot. He finds
a site, usually from links embedded in web pages, then loads
the text from the first page.
He looks for the META tags and pulls out the keywords and
description. If not there he takes the first 200 or so
characters of text and uses them as a description.
The Title is extracted.
He extracts the pure text from the page (strips out the HTML
coding). He takes out the common words leaving what he feels
may be keywords. (Most do not do this last step.)
He now extracts the hyperlinks collating them into those that
belong to this website and those that don't (He visits these
later as this is how he finds new websites).
He may do the same with the email addresses.
He goes on to the next page and so on until he has visited
all of the pages in your web site.
Now he stores all of this information.
He now knows how many pages you have, how many 'outside
hyperlinks in your site', and can give your site a score based
on how it is set up. These are the basics.
What do they do with the info? When someone comes to search
a phrase or keyword, another search routine program takes
over using the information the robot found. A person types in
the keywords and the search program returns the 256,000 pages
matching their keywords. BUT they also consider the following:
How old is the website or how long has the engine known about
it? How large is the website? Was it properly constructed? How
many hyperlinks are there to outside websites?
VERY IMPORTANT! How many hyperlinks are located on other
websites to this site. The older and better the website the
more links to it.
These robots know when you are cheating. You can't trick them.
It is so simple for the robot developer to incorporate code to
negate the tricks. What about scoring keywords only once or
twice per page or area like meta, title, etc? Is this page
close in size to all the other portal pages? How many web pages
in the same directory have the word "index" in them?
Does this site have a lot of content? Is any text the same
color as the background? Are there links to outside sites?
Each page can be checked and compared against what the robot
feels is a statistically normal page. These are computers you
know.
You need a lot of pages with normal content. Instead of
spending the time to make fake pages, give the real ones
content. This will also give your visitors something to come
back to. CONTENT.
Content has been reprinted with permission of the author.
First appeared in www.cyber-robotics.com,
© 1999-2000 David Notestine, all rights remain with author.



Reblog this post [with Zemanta]
Post a Comment