Google ignoring long standing 301 redirects?

[hannamyluv]

I have a client whose back links in GWT went up BUT the links are showing from pages that are 301 redirected from a subdomain and keyword domains. These redirects have been around for months and in some cases years.

I have checked and checked, and the 301's are in place correctly (checked code, checked header, even mimicked a spider and the 301 works and is in place everytime). No changes were made to the site that should have affected these 301 redirects.

Right now, no joke, their link count is up to 77 million backlinks from these supposedly 301 redirected pages. They have lost rank, which is to be expected when 77 million backlinks suddenly appeared out of thin air in less than a week.

Why is Google counting links from pages that technically do not exist?

[tedster]

Sounds like is may be related to the chaos netmeg reported here: [webmasterworld.com...] In her case, Google recently began ignoring the canonical fix 301 redirects.

Google has a technical tangle going on right now. I don't know what we can do except verify that our sites are technically correct, and then wait it out.

[edited by: tedster at 9:33 pm (utc) on Oct 28, 2010]

[scottsonline]

In the last few days google has come looking for pages permanently redirected in 2005. There are no known links pointing to them, so I suspect Tedster is right.

[goodroi]

i agree with tedster.

but there is a small chance it is another issue. if you have poor hosting, it is possible that your hosting company did not properly handle the htaccess file when google visited the page. it is not likely but i have encountered it in the past. i doubt this is the issue just sharing it for sake of thoroughness.

[hannamyluv]

I now have 3 diff clients (diff developers, diff servers) that are seeing this happen. I can't be the only one unlucky enough to have this happen. Anyone else seeing anything like this (backlinks in GWT shooting up from 301 redirected pages combined with a loss in rank and traffic)?

[netmeg]

I don't have poor hosting. I host on my own servers. Have since 1996.

I've got all kinds of goofy stuff going on I don't understand. One site showing www for the home page when that's redirected. Another site suddenly showing https for the home page when it's been properly redirected (and previously indexed correctly!) since it launched in April. Yet another site (one of my own personal sites) has a particular PR4 page with plenty of back links dropping in and out of the index like crazy. It's actually gone UP in TBPR, but is currently not indexed, as far as I can tell. That one's had a redirect since 2006. I even took off the canonical tag just in case that was screwing it up, but it didn't help. (This particular one is not a new situation; it's been bouncing around for about three months now. No idea why)

There are a few other sites with oddities as well. None of these are brand spanking new sites, and all of them were indexed correctly at one time. So it's not a question of going in wrong and not getting fixed. They were in RIGHT and now are broken.

[tedster]

Just to be clear - these URLs are showing in the WMT links report, but not in the SERPs - correct?

[hannamyluv]

Mostly, but for the first client who had an issue that included keyword domains, a few pages are showing in the SERPs for the keyword domains. No cache and they keyword domains have the "youngest" redirects.

Mostly the issue is located in GWT. The thing that is boggling my mind is even if they have counted every link on every non-existant page, where the heck they are getting 50+ million links. There should not be that many.

And just to be clear on this, these are not small site. These are not small businesses. These are decent size companies with sites that get pretty decent traffic.

[netmeg]

No, they're showing this way in the SERPS.

[tedster]

Sometimes linking sites, especially automated "black hat" sites, have major canonical issues and "infinite crawl spaces" that can be indexed, dramatically inflating the backlink count. Google now seems to just do a data dump into the WMT link report and these don't get filtered out.

That's my theory and I'm sticking to it ;)