I would like to know the number of links there are to my Web site. 

When I use Alta Vista, I get unreliable results. For example, if I
enter "link:http://www.ford.com/ -host:ford.*" to find all of the
links to Ford except those within Ford's domains I get 39,975. If I
open another browser and enter the same search I get 43,030 results.

When I use Google I get relatively low results. The same search for
Ford (without excluding the domain) returns 7,940 results.

Where can I get the most accurate count of links to my site? What
should I enter as a search term? (I would like to know the number of
links to my site excluding links from my own domain.)

Hi, 

While Taxmama's comment below is an example of other sources to get an
idea of who is linking to your site it is not quite accurate. As if I
am on this page, and I type into the address bar your website and hit
the enter key, you logs will reflect that answers.google.com is
linking to your website, which of course is not the case.

Your question is difficult to answer, and certainly can not be
answered by search engine. Both Yahoo and Google and all the other
search engines recognize a link to a site in various ways. The highest
accuracy is going to come from a webrobot you have made that can go
through the Internet and check for itself as to how many and where
links are being created. Even this is not going to be 100%, for
instance, a link there today on a dynamic page may not be there in two
weeks time or even by the time the robot has completed it's circuit.

Google's number is much lower for a reason. They are showing you links
that they consider "good links" for PageRank. If a page out there is
linking to you, but has no, or a low, PageRank, it won't show up on
the list of "good links" to your website. Yahoo's maybe more accurate
but will list the same link several times from a web domain that links
to you from several pages on their website. Something Google tries to
filter most of the time. Also, it won't take into account the fact
that the page may be dynamic and a different name this week. So, you
have two links which are really one. This is not really an accuracy
problem for Yahoo, but more of a resource problem.

The creation of your own webbot is rather simple and can be done by
any competent Perl programmer or C++ programmer that has experience in
network programming.

For information on webbots you can look at this link here. 
http://www.searchtools.com/robots/robot-code.html


Thanks, 

webadept-ga

---

Request for Answer Clarification by spikedonner-ga on 02 Dec 2002 10:50 PST

I understand that there are different ways to define a "link" but that
doesn't explain why one site give completely different numbers for
requests within seconds of each other.

Since a webbot is not an option for me (or for many folks who don't
own their own ISP) what is the best SEARCH ENGINE to use and what
should I enter as a search term?

---

Clarification of Answer by webadept-ga on 02 Dec 2002 12:42 PST

Hi, 

The one I use to "get an idea" of the amount of links is this page
here:

http://www.linkpopularity.com/

It requests from three search engines at the same time. The reason the
counts could vary so radically is that search engine databases are
never static and are continuously updated with information. I've seen
changes up to a difference of 20-25 links in 15 minutes of time.
Reasons vary from the page wasn't available at the count time, the
link was broken, the link was changed, the page was changed, etc.
Search Engine systems are really nothing like most databases you are
familiar with. Their size is another problem. Also, the databases are
on several (20-30) different computers, so your first request could
have been run on one server, and the second on a completely different
server, which has been updated at a different time.

Most of the time (95%) Google's count will be less than the others,
because of the way they see "pages" and vaild links. If a site's
PageRank is less than 3 it is not likely that they links on those
pages to other sites will be recorded. This is something to keep in
mind if you are purchasing links on other sites to boost your own
PageRank. Make sure that the page you are getting a link on has a
PageRank of at least 5, otherwise you are just wasting your money.

Thanks, 

webadept-ga

Good description of how link rankings are derived.

Hi Spike,

Do you really, honestly truly want to know the absolute accurate number
of sites linked to yours? The ones that are real links, not free for all
links and other garbage? The ones that people actually use, at least on
occasion? 

Look in your site's log files. When anyone visits your site from
another one, it's recorded in your own log files. 

If you need to see them in a more visual format, ask your webhost
to provide you with the tools. 

Checking in search engines? That's not accurate. 
Besides, they often include dead links.

Incidentally, just for fun, look at alexa.com

Best wishes,

Your TaxMama-ga

P.S. Three of sites AltaVista shows linked to mine are bad links,
on the first 3 pages...

"As if I am on this page, and I type into the address bar your website
and hit
the enter key, you logs will reflect that answers.google.com is
linking to your website, which of course is not the case."

This information is incorrect.

in usch a situation the log files will record a direct hit (user
entered the URL themselves). A referrer is only passed to the web
server if a link on a page is clicked.

I just though I would correct the inaccuracy.

Hmm.. Then CNN and Disney have links to my website.. how interesting.
Perhaps you should check this, because according to the logs on Apache
this is not the case.