Make *all* your links count!
by Randall McCarley
November 7th, 2006
It has come to my attention that I could be doing a better job with my copywriting on this blog. For example, when I wrote about the Alexa interface update the other day and then Donna did the same her title - Alexa has a sexy new traffic rankings UI - was much better.
Last month I announced a tool that solves the canonical issue with the search engines. Of course most people don’t know what that means… even SEOs. A better approach may have been something more like…
We all know links act as votes when it comes to search engine rankings. But what do you do when a search engine only counts half your links? Odds are, this is happening to you!
The search engines see two versions of your site. The first is the regular domain.com and the second is your domain with the “www” prefix. When some people link to your site with the prefix and other people omit it both versions send traffic to the same site but the search engines see it as two different things!
This is called the canonical name issue and it affects most sites listed in the search engines. Using Google’s nifty define: operator I got this definition:
The real name of a host. Used in CNAME records, PTR records, NS records and MX records. A Canonical Name is something of a fiction because many servers have more then one equally valid name. Basically, any domain name that has an A record.
Most domains come with two A records that are equally valid. The first is the regular “domain.com” type the second is with the www-prefix like “www.domain.com”.
You can check to see if your site is affected by the canonical issue by typing it directly into your browser URL bar (example poetry.com). Then type it with the www prefix (example: www.poetry.com). Notice how the “www” part is there is you type it in explicitly or not of you don’t? That’s the problem.
Big companies and SEOs have been aware of this for a while and most big name sites have made the correction:
| Without www | With www |
|---|---|
| Microsoft | Microsoft |
| 14th Colony | 14th Colony |
Notice each of these redirect to the www version regardless of what you enter in the URL bar of your browser.
As confusing as this can be - even the engineers at Google, Microsoft and Yahoo are having a hard time fixing it - there is a simple solution.
By typing your domain name into the Seeing Double tool the code you need to upload to your site is automatically generated. This removes the possibility of errors on your end and makes fixing this problem take just a couple minutes.
PS. Special thanks to Wit who got me thinking about copywriting in this thread about .htaccess
Next Article: Fast-Acting Pops Link-baits the Refuge during Down-Time Previous Article: I need volunteers for a Google sandbox escape experiment
November 9th, 2006 at 5:03 am
What extension must the .htaccess file have. I pasted the code in notepad.
.txt, .htm?
One more question, I read about the social bookmarking sites. I am interested to know about some that allow importing from delicios bookmarks.
Thanks!!
November 9th, 2006 at 4:55 pm
.htaccess actually is the extension in a strange way. The peiod is in front with noting in front of it! This is to make the file “hidden” on linux systems (servers) but is part of the protocal.
I know Furl and Blinklist will let you import del.icio.us bookmarks. I’m sure some others do too.
June 16th, 2007 at 11:38 pm
I have added this to my .htaccess. Nice! I always wondered how the others where doing it??. Thanks.
However when I ported this code some of my images I hotlinked to (on my OWN servers) from my other sites vanished. I had to copy images and host them right from under the domain. Any remarkes?
June 16th, 2007 at 11:49 pm
If you have virtual domains hosted under your main /root/ they won’t work if this code is installed in the top /root/ . Had to turn it off.
June 17th, 2007 at 12:38 am
Sorry to hear that but it sounds like a server setting on the host side. I use virtual hosting and haven’t had a problem. Also, I don’t see how this .htaccess code would affect images no matter where they are hosted. It only repoints one of the domains to the other. I’ll see if someone can help you out with this.
June 17th, 2007 at 9:43 am
Ok. Here is what I found out through trial and error: If you have virtual domains under your main /root/ you need to add the code in to the virtual /root/(s) as well and just set to correct domain for virtuals. That part works.
As far as hotlinking images, I dont know. But not a big deal
June 17th, 2007 at 10:00 am
Well cool. I still have a couple feelers out there to see if someone knows the answer.
June 19th, 2007 at 10:46 pm
I’ve never had image trouple. My guess would be that if you’re adding www to the URLs, the image code doesn’t have them, and you got the whole URL in the image code.