4

Google appears to have two documents for Soft 404:

https://support.google.com/webmasters/answer/181708?hl=en&ctx=tltp

https://support.google.com/webmasters/answer/2409443?ctx=MCE&ctx=S4

Neither one goes into any detail of what criteria is used to qualify a unique 200 OK page under their "Soft 404" term.

Is there a certain number of words or characters that are minimally required to make a page unique, and bypass a "Soft 404" qualification?

Improve this question
13
  • There is another document about it here too: googlewebmastercentral.blogspot.com/2008/08/… – Stephen Ostermiller♦ Dec 23 '14 at 1:13
  • 1
    I would not worry about Google's notion of a soft 404. Most of the time it is a simple phrase in the content such as not found or even a longer phrase like cannot be found until morning which you would think would not qualify. There is a fair number of these phrases that Google uses. Some pages, it seems lately, are tagged with soft 404 errors that may have less content. I have some pages that I would not call thin content get tagged as soft 404s. This is nothing to fix. It is a Google silliness. Look for a phrase that can be changed or expand the page content if you can and relax. – closetnoc Dec 23 '14 at 1:16
  • 1
    @StephenOstermiller What kills me about Google's soft 404 is the automatic assumption that that is what is going on. They assume you are sending them a soft 404. Google kills me in that it cannot understand that it takes perfectly valid pages and marks them as soft 404's. I do not have soft 404's on my site and yet Google has 55 of them. I have found them recognizing the most bizarre phrases because they do not look for not found but not and found within proximity. Also, lately it seems to be marking smaller but perfectly valid pages as soft 404's. They are driving me nuts with this. – closetnoc Dec 23 '14 at 1:24
  • @closetnoc, no, i don't have any not and found on the pages that google thinks are Soft 404. it looks like size appears to be the primary determination in my case. – cnst Dec 23 '14 at 2:22
  • 2
    If Google is getting soft-404s wrong for your site, send me some sample URLs and I'll forward them to the team here. You shouldn't need to artificially tweak the response size. – John Mueller Dec 23 '14 at 14:02
3

There must not be a possible word limit for a particular page to identify it as soft 404. The ideal cases when such issues are reported by Google are:

  1. When a dynamic template with no content (due to some incorrect parameters) returns 200. In such scenario, there can be thousands of pages which might not be important for anyone, but are getting indexed because of some incorrect query parameter handling.
  2. When a real page is redirecting (generally 302) to a 404 page. In such case, rather redirecting to a 404 custom page, the page itself must either show 404 or 410. Its a general misconception to redirect to a 404 page.

Just ensure the pages that are reported are fine and have worthy content on it for a user.

Hope this helps!

Improve this answer
3
  • No, neither one case is applicable to my situation. I have absolutely unique (but small) pages on a couple of distinct websites that are reported as Soft 404 by Google Webmaster Tools. I'm not doing any redirects to a 404 page, either; besides, a 302 to a 404 is unlikely to be Soft 404 -- it sounds like it should be a normal 404, and it most likely is, as I've seen no evidence to the contrary. – cnst Dec 23 '14 at 6:01
  • Ok. If it doesn't violates your privacy, could you please share some URL's which are reported as soft 404 (atleast 4 of similar type) and also the custom 404 page of your website. – keepitup.sid Dec 24 '14 at 18:16
  • they've already been shared in the neighbouring answer; i don't actually have any custom 404 pages yet, nginx replies with the default pages and default error codes – cnst Dec 24 '14 at 18:36
1

There appear to be several criteria that can qualify a page as a soft 404:

  • Redirect to the home page
  • The phrase "not found" or equivalent in a prominent place on the page
  • A blank or nearly blank page

I don't know exactly how much text you need on a page for it to be considered non-blank, but that is certainly not the usual reason that pages are soft 404. I tend to get them for redirecting to the home page. Many sites are configured to output the text "product not found" on a 200 page when a lookup in their catalog fails.

The page that you mentioned in the comment looks somewhat like an Apache error page: (image source)

My guess is that it isn't just that your page isn't large, but also that it mentions the word "Apache".

Google's John Mueller says in the comments:

If Google is getting soft-404s wrong for your site, send me some sample URLs and I'll forward them to the team here. You shouldn't need to artificially tweak the response size.

The best way to contact him appears to be through Google+.

Improve this answer
5
  • I dunno about other sites, but my sites certainly don't have any of the scenarios you describe, yet my Google Webmaster Tools is now filled with all of these Soft 404s. Your answer doesn't really address my question at all. – cnst Dec 23 '14 at 2:25
  • @cnst You have John's attention on this matter. There is no better contact in Google for this. I suggest contacting him and working with Google directly to solve this issue. It will help all of us who have seen this oddity lately. Thanks in Advance! – closetnoc Dec 23 '14 at 15:07
  • @closetnoc, if only there was a way to contact him! – cnst Dec 23 '14 at 15:20
  • @cnst I see what you mean. Short of a Google+ account, I do not see too many options other than posting some URLs in the question and using @ here to contact him that way. Once John has the URLs, then you can re-edit your question to remove them if you want. It's up to you. – closetnoc Dec 23 '14 at 15:33
  • The ports.su/www/apache-httpd-openbsd looks nothing like your standard 404 Not Found! But, actually, a good guess! :-) Unfortunately, there are many other pages that are likewise reported as being Soft 404, so, the words apache are unlikely to be the contributor. However, looking at the list (16 pages out of ~10k are "Soft 404"), it does appear that they all reference some kind of web-based thing in their URL -- ports.su/www/pear-HTML-Select, ports.su/www/p5-HTML-StripScripts-Parser, ports.su/www/p5-WWW-YouTube-Download, ports.su/net/pear-Net-URL etc. – cnst Dec 23 '14 at 15:39

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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