Google says my soft 404s are increasing - am I doing something wrong?

[denisl]

I have been trying to reduce the number of Soft 404's reported in WMT for a couple of years, since taking a knock in Google.

These are dynamic pages where if the amount of information displayed is below a certain level, the following is now inserted in the html head:
<META NAME="ROBOTS" CONTENT="NOINDEX">

Why is Google listing this as a soft 404 - am I doing something wrong with the code - I have stared at it looking for a spelling mistake, just incase. These pages may be thin - might contain some sort of warning message, but are not true 404's.

The number listed started dropping a few weeks ago but is increasing again now

[Wilburforce]

Google will flag up a "Soft 404" when the server returns a code that is not 404 for a URL that does not exist. Redirects - e.g. redirecting requests for non-existent pages to a custom error page or the Home page - are a common cause. It makes no difference whether or not the page that is actually returned is indexed.

Their own Help section might be useful: [support.google.com ]

If this is happening with dynamic content, there is probably a redirect (if somebody clicks on a link to THIS URL, go and get THAT page) somewhere in the process, and you will need to find and change what is happening there if you want to eliminate the Soft 404s.

While Google at least implies disapproval of Soft 404s, it isn't likely to have much impact on SERPs, and if your site is slipping in position there are probably other areas that are more important to look at.

[denisl]

Thank you for the response. My aim was to remove thin pages from the index and avoid a penalty. Originally the soft 404's listed by Google were all thin pages (which have a purpose), which were in the index. Having noindexed them, I didn't expect Google to continue to flag them as soft 404's.

[aakk9999]

Google Webmaster Tools (Search Console) has known to be buggy. Maybe you need to give it a bit more time for catch up.
In a meantime, as a very crude signal, you could perhaps do a combination of site: and inurl: commands where inurl would have a string that is unique to all these pages you have noindexed. Over coming weeks you should be able to see the number of results to this query dropping - indicating Google is removing these pages from its index.

[lucy24]

Take a quick look at a few days' raw logs. When google requests one of these pages, what response do they get?

[denisl]

They will get a 200 response ( it isnt a 404 - there is a small amount of content), but having noindexed it for the last 2 years, I didn't expect G to report it as a soft 404. I see a soft 404 as thin content and a potential penalty.

[lucy24]

Does your site ever, under any circumstances, return a 404? Google tests by requesting a garbage URL like hgmjkhtghyfgkyh.html that can't possibly be a valid page. (This appears to be programmatically triggered any time a site starts returning an anomalous number of 301s. I see a flurry of them periodically when I've rearranged a directory.) If even this gets a 200 response, there's a problem.

The "noindex" tag is really unrelated to the "soft 404" issue; at this point they're testing whether the URL exists at all.