Seek help correcting large number of 404 errors generated, 95% traffic halt

[mantucket]

Hi, The following GWT screen tells a bit of the story:

site: http://bit.ly/mrgdD0

http://www.diigo.com/item/image/1dbpl/wrbp

On about Feb 8 I decided to fix a large number of 'duplicate title' warnings being reported in GWT "HTML Suggestions" -- these were for URLs which differed only in parameter case, and which had Canonical tags, but were still reported as dups in GWT.

My traffic had been steady at about 1000 clicks/day.

At midnight on 2/10, google traffic completely halted, down to 11 clicks/day.

I submitted a recon request and was told 'no manual penalty'

Also, the 'sitemap' indexes in GWT showed 'pending' for 24x7 starting then.

By about the 18th, the 'duplicate titles'  count dropped to about 600 or so... the next day traffic hopped right back to about 800 clicks/day - for a week - then stopped again, down to 10/day,  a week later, on the 26th.

I then noticed that GWT was reporting 20K page-not found errors - this has now grown to 35K such errors!

I realized that bogus internal links were being generated as I failed to disable the PHP warning messages.... so I disabled PHP warnings and fixed what I thought was the source of the errors.

However, the not-found count continues to climb -- and I don't know where these bad internal links are coming from, because the GWT report lists these link sources as 'unavailable'.

I'v been through a similar problem last year and it took months (4) for google to digest all the bogus pages ad recover.  If I have to wait that long again I will lose much $$.

Assuming that the large number of 404 internal errors is the reason for the sudden shutoff...

How can I a) verify the source of these internal links, given that google says the source pages are 'unavailable'..

Most critically, how can I do a 'RESET" and have google re-spider my site -- or block the signature of these URLs in order to get rid of these errors ASAP??

thanks

[randfish]

Hi Mark - wow, sounds really rough. I've got a few suggestions:

• First off, you need to make 100% sure that you've actually fixed the issue and that the internal links are pointing to the right places AND any old URLs that may have had internal/external links are either rel=canonicaling or 301 redirecting to the correct, updated locations.
• You might try using a few tools to verify this, including the SEOmoz Crawl Test http://pro.seomoz.org/tools/crawl-test and Screaming Frog: http://www.screamingfrog.co.uk/seo-spider/
• When you are ready, submit new XML Sitemaps to Google with the proper URLs. Make sure you've deleted/removed your old ones.
• You can also send the reconsideration request again, indicating that while you're aware this isn't a penalty, you have realized some technical/navigation issues on the site and believe you've now fixed these.

Hope this helps and wish you the best of luck!

[iPullRank]

I agree with Rand's suggestions. I just ran a Screaming Frog crawl of the whole site on 10,233 links, 8997 URLs and got no 404s. So I think it's pretty safe to assume you've fixed the 404 issue. Here's the output of the crawl in case you'd like it for a reference: http://www.sendspace.com/file/7zui0v

I'd say:

• Definitely clean up and resubmit your XML sitemap
• Double check your backlink profile with Open Site Explorer and MajesticSEO to be sure that there aren't sites linking to URLs that no longer exist. If you find any of these make sure to 301 redirect them. Just take all the target URLs and dump them into Screaming Frog in list mode. All the links from OSE point to your homepage so they are not an issue, I don't have access to Majestic right now so I couldn't run those for you.
• You can now Submit pages in Google Webmaster Tools as well in the Fetch as Googlebot section. So you may consider submitting some of the new pages the site generates in addition to your reconsideration request to help get Google to re-crawl and find the 404s are gone.

Good luck man and please let us know if nothing changes after you implement these fixes.

-Mike

[mantucket]

Thanks for the replies guys - - I had run Xenu on the site and it found no broken links... but still GWT error count continues to climb, and as of today

Google released a MUCH improved timeline view for the error count --- problem is, it's still showing 58K errors as of yesterday and climbing, long after I fixed them - and it wont show me where it thinks the source is...

These errors are all on internal pages BTW..

Heres the new google view

http://awesomescreenshot.com/0ef1gy6c7

The new GUI also includes a way to mark errors 'fixed' -- one by one!! I need to mark 60 thousand at once!

Also I can see the date these errors started appearing and it just doesnt make sense given that is the day my traffic started reappearing as well..

[randfish]

I wouldn't feel too confident that the numbers and dates Google's showing you are precise or accurate. In fact, we've seen times when GWMT is considerably off. I'd watch how Google crawls your site and look at search traffic to your pages - those are likely leading indicators that things are/will be fixed.

[TheBigK]

Hello Rand, I've been facing a similar problem with my site. I'd really appreciate your response here - http://www.seomoz.org/q/help-fixing-the-traffic-drop-that-started-on-4-september-2012.