|Relative vs absolute links in the new Google|
trying to figure this out...
If I were truly an advanced web professional I would probably know what's going on right now with my internal links and Google; I'm not, so I don't.
I have great problems in Esmeralda with www.domain.org vs domain.org. By and large, my site is doing well, but I have now 102 pages as www.domain.org and another 50 as domain.org. My index is gone, and it was one of the few pages that had a mix of absolute and relative internal links. My main page has one incoming www.domain.org backlink, and a lot of relative internal backlinks, and this seems to have confused Google. Several other pages have a mix of absolute and relative internal links and they are also showing a messed-up mix of www.domain.org vs domain.org. I'm very confused because none of the other SE's have had a problem with this and Google never had until now.
I'm wondering whether the internal links have to be one or the other, absolute or relative, for any particular page. I never realized this could be a problem until Esmeralda. Any advice or input would be greatly appreciated.
Edit: Changed "backlinks" to "pages".
Google, like any browser client, will fully resolve relative links before it uses them. So relative links can't "hurt" you.
However, if Google finds external links to "the wrong domain", then what you lose with relative links is the ability to "out-vote" those incorrect external links.
However, if you 301-redirect all alternate domain names to one single domain name, then that is is not a problem, since "incorrect" domain names get corrected by the redirect before any relative link resolutions are attempted.
Advice? - Wait a week. If there are still problems, then worry. It is apparent from other posts that Google has not had time to go through their index removing and consolidating duplicate domains. When they do this, I suspect your problem will go away. In the meantime, you might want to search for the many posts here on how to 301-redirect multiple domain names to one "standard" domain name to prevent this problem from happening again.
On one of my "hobby" sites, I redirect all requests for mysite.com, mysite.org, and www.mysite.com to www.mysite.org. I have never had the problem you are describing, despite the fact that most of my internal links are relative, because I *specified* which domain was to be used, without leaving it up to Googlebot.
You know, this is odd because I just looked at some sites in my category, to see if there was a different between somedomain.tld and www.somedomain.tld.
I found nothing.
Having read up on others' experiences here, I did rewrite .htaccess to redirect www.mydomain to mydomain, and I'm happy I did. But it does seem like Google's problem with www vs. not is not an across-the-board problem but one that's only happening to specific sites.
I looked at link:fairlypopularsite.tld vs. link:www.fairlypopularsite.tld. Both showed exactly the same backlinks (there was only ONE link on those sites, although I didn't check which one it pointed to). I also checked link:www.geocities.com/somesite versus link:geocities.com/somesite. Again, both were identical.
I wonder why some sites are running into serious problems with this, whereas in the cases I was looking at, Google seems to have properly realized that www.foo and foo are exactly the same.
Many thanks JD.
|Google, like any browser client, will fully resolve relative links before it uses them so relative links can't "hurt" you. |
That's what I've understood, and have been working with, but I seem to be finding many pages listed in Google, as domain.org/page, that have only internal relative links. I haven't noticed this until Esmeralda. I keep wondering if it's because of the few pages that have a mix of relative and absolute internal links. I only have one incorrect external incoming link that I can find, a link to the index that comes in on domain.org. I'm at a loss as to what is going on. It seems to have eliminated my index page somehow.
Your advice to wait a week is probably best. Thanks again.
A lot of our sites have relative links back to the homepage from the internal subpages and G has never had a problem sorting this out...nor are they now. We've done it that way just because it's a bit of a time saver in dev versus the hard coding.
What we have seen, however, are problems in this update with external backlinks coded in different variations...
Just a follow up thought. Is it possible that Google's not having sorted this out yet is a significant cause behind the SERP's being a bit wacky, and that when they recombine everything so that homepages get their full weight, those "lost pages" will reappear, and the index will all get a lot closer to where it should be? Couldn't be that simple, right?
|A lot of our sites have relative links back to the homepage from the internal subpages and G has never had a problem sorting this out |
Could it be that those sites are consistent in their internal linking and mine is not? I only have a few pages that are internally linked to by both absolute and relative links, and have been trying to find them and make them consistent since yesterday.
I really don't think that "consistency" on relative/absolute links should matter very much. I think it is simply that you allow your site to be accessed by either domain, and that there *are* external incoming links that point to the incorrect domain. If Google enters your site using the "incorrect" domain, then every time it resolves a relative link, it will resolve it to that "incorrect" domain, thus adding "weight" to the incorrect domain.
I'm not a day-to-day update-follower, but my perception is that the last two updates have been very strange - I suspect because Google is making *big* changes to both the search programs/algorithms and to the databases on the actual search machines. Because of this, there may be some problems that won't get handled until all the big changes are settled.
As far as preventing this kind of thing, I'd say that "neatness counts" as far as making sure you don't have a bunch of different (unredirected) domains floating around, problems in robots.txt, incorrect or missing cache-control headers, and that kind of thing. Or maybe it's just me - I tend to be "detail-oriented" :)
Google does not want a bunch of duplicate domains cluttering up their SERPs and for this reason, I'm certain they will resolve the problem as quickly as the mechanics of the current update allow.
As one who suffers this domain.com vs www.domain.com issue. I, at first, thought it was because I have 187 links to domain.com by other sites (their error). But I see here where this is happening when there is only 1 inbound link to the wrong domain. So, that kind of blows up my theory. Also, someone said, it might be that google enters on the wrong domain through a backlink and then indexes it relative to that. That makes sense, too. Except that both domain.com and www.domain.com for my site are in all datacenters. The problem is that the index page for www.domain.com (PR6) which has far far more backlinks is buried in Serp's, while domain.com (PR? at this point, saw PR4 then PR0) is showing up first in SERP's. Seems the higher PR domain would rank higher.
The only optimism I have seen in this scenario is that -fi recouped my site back to it's #1 position for kw's. I think this indicates that they are wroking out this duplicate content filter issue, and we will see the end of this strange issue when this new index is fully baked.
Until then, I sit and watch, with bated breath, -ex and -cw where www.mydomain.com is again buried.
|Advice? - Wait a week. <snip> |
> Until then, I sit and watch, with bated breath, <snip>
More advice: Don't sit and watch. If you're not spamming, you've got nothing to worry about. It is in Google's best interest to "get it right" and worrying about it will get you... nothing but worry.
I has been noted by GG that there is a distinct lack of senior member posting in the update thread. It's not because of "snobbishness," it's because most have BTDT and got *tired* of/by it. :) Watching this update - other than just checking in once in awhile to see how many datacenters are updated - and responding to each little change with worry is like going into the auto repair shop to check on your car (which is in to have the engine rebuilt), and panicking because it won't start and you can't drive it, even though the engine block has been removed and is obviously sitting in a cradle while a mechanic reinstalls the top-end.
I hope everyone can stop worrying about this until at least the middle of next week. At one datacenter update per day, the datacenters should be updated by tomorrow at this time or maybe Sunday. Googleguy has stated that there will then be some further filtering or tweaks, and maybe additional backlinks added in after that. So next week, we may be able to begin to analyze things. If sites are still missing or scrambled then, I for one will be very surprised. I suspect this update is rocky because of big changes, and that future updates will not be anywhere near as turbulent. Some are guessing (because of comments by GG) that this is the last update before Freshbot becomes the only 'bot, and the index will be updated constantly from now on. I think that would be worth it -- after all, Google must not enjoy the fact that the AltaVista index bas been fresher for the last 2 months or so! Also, it makes "tuning-up" your sites easier if they update frequently... And... if you were Google, wouldn't you want to get this kind of major change done before you had stockholders to answer to? (Just a guess)...
Before anyone flames me, let me say that yes, I've got people calling me worrying about "plummeting" to (Saints preserve us!) #3 in the SERPs, and thinking that is going to destroy their business, etc., etc., and it's somehow all my fault - Not! But I've learned to deal with it because it's too tiring to worry about it. Plus, those sites are clean as a whistle, and they're good sites, so why worry?
Please y'all, get some rest, and don't worry. I think it's going to be "worth it" in the long run. Those are some fairly competent people out there in Ca.
|I think it is simply that you allow your site to be accessed by either domain, and that there *are* external incoming links that point to the incorrect domain. |
I think you've got it figured out there, it comes in via that incorrect link, follows internal links... It really surprises me that one incoming link that has no www on it can cause all these problems, but so it goes.
I've asked the person linking to me to change the url, he says he will, but he lives in the jungle, (or rather quite close to it), does his email through a cellphone, and only updates the site when he gets to a landline near "civilization". This could take a while...
In the interim, I'll keep adding pages that still have to go up, and hope that Google sorts it out.
I also have a site that suffers fromthe domain.com vs www.domain.com problem. It's a fairly common problem. You can see it in action in the Google directory. Just go through some areas you know well and look at the sites with no PR, at the bottom of the lists. Some of these sites will actually be rather popular sites with lots of links to them, but the reason they're showing no PR is that ODP is failing to use (or is using) www. when most of the other links don't.