Webmaster Tools is showing thousands of 404 errors, where pages on the site are referring to another incorrect URL. For example, URL not found www.example.com/shop/=, linked from www.example.com/shop/gift-voucher and www.example.com/shop/special-plant-offers.

I obviously have checked the source and cannot find any references to this link on any page.

The only consistent issue is that it only seems to report this error on pages with two section i.e. www.example.com/shop does not report any error whilst all pages with www.example.com/shop/xxx (where xxx can be several different pages such as gift-voucher) all report this.

I cannot seem to duplicate this error. I have run a link checker (we use Screaming Frog) and it does not report this error. I have fetched these pages as a bot, and these do not report this error.

I am at a total loss. I cannot even duplicate the issue, but it is most definitely an issue, as Webmaster Tools is reporting new errors every day.

Is this perhaps Google bot doing its own thing?

  • Take a look at this question from Google's forum's may have an answer support.google.com/webmasters/bin/…
    – Anagio
    Dec 5, 2012 at 17:10
  • Unfortunately the generic google page on 404's does not really help (and this issue is way more complex than that addressed in the google help page). I could repost this question on google's forum, but I prefer stack exchange
    – C4PO
    Dec 5, 2012 at 17:30
  • Couldn't be Googlebot trying to execute some JavaScript, could it? I've had similar woes and that's occasionally looked like the culprit.
    – GDVS
    Dec 5, 2012 at 18:08
  • 1
    @plantify i'd repost it there as well, it's not in the SE network so it's perfectly fine to post it there as well. Especially since there are Googlers chiming in on the forums once in a blue.
    – Anagio
    Dec 5, 2012 at 18:15

2 Answers 2


I cannot verify this authoritatively, but based on my own observations I'm going to hypothesise that once Google has cached a page, it will continue for some time to follow links from its own cached version of the page rather than from the live version of the page.

As far as I can see, the two quickest ways to get around this are:

1) to re-fetch the live version of the page in Google Webmaster Tools, so that the Google cache has an up-to-date copy of the live page; or

2) to update mod_rewrite rules in .htaccess, so that when Google tries to follow an outdated link to a non-existent location, it is automatically redirected by the server to an existing location.


Two thoughts

  1. Can you enforce a trailing slash on your URLs. For example, www.example.com/shop/special-plant-offers/ (Sorry, it's 2:30 am and I just forgot the explanation for why this might work.)
  2. The equal sign "=" looks suspiciously like a special character to me. A) I have had a few problems with copy and paste code containing weird unicode that was invisible to the naked eye. Typing the URLs or code by hand fixed the problem. B) Is your file actually encoded as UTF8 w/o BOM? C) Does your <head> declare the proper charset? D) Does your server's response header properly identify the actual charset of the served document?

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.