Could your error pages affect your Google ranking?

[bumpski]

Is there any chance Google is purposely crawling website error pages, looking at what's returned, and using it as a factor in website search results ranking? Anybody have evidence of this?

I'm sort of a believer in the KISS principal, and as a result, at least in the last couple years, have relied on the Apache server default error pages. They are nice and simple and I'm sure most visitors are familiar with the format.

Well my webhost has hijacked my nice default error pages (yet again without my knowledge). Not only that! They've enabled wildcard subdomains. Why? Well they're slamming IFramed ads all over the error pages and actually embedding my domain names at the top of these pages and Titles, etc. In addition the advertisers they are displaying are popping up more browsers with ads, etc!

So has anybody noticed Google intentionally crawling error pages and perhaps evaluating what is found? Right now for me it could have serious impact. I'll fix it pretty quick but in the last couple months I've seen some traffic degradation from Google and wondered if this could be a factor.

Should I dump this host, I'd love to. Only they didn't used to be my host until they bought out my host, a host with 700,000 sites. This big conglomerate is snapping up hosts as rapidly as it can, at least 40 a couple years ago. It masks these buyouts and migrations as "Control Panel" feature updates! This superhost is now so big if they enable wild card domains, millions, yes millions of sites, may be loading up Google with duplicate content. Millions of sites may just start showing pop ups on their error pages unbeknown-st to them. So if I do relocate how do I avoid this webhost Pacman? I don't think a power pill will do it!

[goodroi]

anything is possible since google performs hundreds of tests and is making constant tweaks to the algo.

i would doubt that error pages on their own are a significant (if any) part of the ranking formula. what is more likely is that a site has many broken links which then points to an error page and thus google devaluing that site.

based on my experience backlinks, content and usage signals are the three main factors in ranking. i would focus on those areas. i would not worry about google poking around error pages i would just make sure that i didnt have any broken links or 404 errors.

[tedster]

One reason that Google may index the content of intended error pages is that the actual 404 status isn't returned in the http header. It's still common to see a 302 response followed by a 200 OK status for the "custom error page". In fact, this handling used to be recommended in some older IIS server manuals.

This is what Google has been calling a "soft 404" and they are sending out emails to some webmasters when they see the problem at a large scale.

[iamlost]

Is there any chance Google is purposely crawling website error pages, looking at what's returned, and using it as a factor in website search results ranking? Anybody have evidence of this?

Chance? Sure :)
Evidence? No.
However, a bare bones server 404 (or any other response code) is not exactly 'user friendly'. I built custom response pages for each of my sites a long time ago in order to increase visitor retention. And it worked/works great for those instances where redirects did/do not apply. If it also gives a SERP boost that's nice, but not why it's worth doing.

Note: I do look at 404s regularly to see if redirection is worth implementing.

Well my webhost has hijacked my nice default error pages...

I would have been gone the next day. And my law-type-person would have been parsing the service contract.

[TheMadScientist]

Personally, I almost always use custom error pages, and I can understand the wildcard subdomains, personally I like having them on, but I'm fairly good with mod_rewrite so it isn't really an issue for me...

I would recommend 2 things to avoid this in the future:
1.) Get some Mod_Rewrite in place to canonicalize / consolidate all subdomains not in use to the main domain.

2.) Always use custom error pages... I like to put a 'mini sitemap' on mine, because it keeps people like me from clicking back and not visiting your site again... No joke, I started doing it because of the way I surf. If you show me two 404 error pages in a row without a link to the home page and main sections of your site it goes on the mental 'don't ever visit' list I won't visit it again.

You can find help setting both up correctly in the Apache Forum [webmasterworld.com]

Make sure you get both of these right, and definitely make sure you serve a 404 error on the 404 error page... I usually use PHP for mine just so I can set the error on the page visitors see as a 'fail safe' in case something 'goofy' happens with Apache or redirection... EG header('HTTP/1.1 404 Not Found'); ALWAYS goes on my custom error pages.

ADDED: Forgot to answer the question...
Sure they could, but I'll let someone else who wants to try to test and see if they do... What they effect more than anything (as noted above) and IMO is visitors. I don't really care to find out what if any effect they have on SEs, because my visitors are what's really important and ugly or unhelpful error pages definitely effect them plus IMO 'unhelpful' error page negatively impact the 'real' (people) visitor perception of the site.

[valex]

Well my webhost has hijacked my nice default error pages (yet again without my knowledge). Not only that! They've enabled wildcard subdomains. Why? Well they're slamming IFramed ads all over the error pages and actually embedding my domain names at the top of these pages and Titles, etc. In addition the advertisers they are displaying are popping up more browsers with ads, etc!

Name them and shame them :) I suspect 1&1!?!

I built custom response pages for each of my sites

This is the best practice and it's working perfectly fine for me as well.

[tedster]

Our normal policy here is not to discuss specific hosting companies in the public forums (Supporters do have a forum dedicated to hosting, however.)

This case is so extreme that we're going to allow an exception to the rule - bumpski is welcome to name the webhosting company involved in this fiasco.

[bumpski]

Name them and shame them :) I suspect 1&1!?!

Please note I've been allowed an exception by WebmasterWorld to name some names in this special case only.

In this case the sites were originally hosted by IPowerWeb and I did have custom error pages. But then unknown to me EIG Endurance International Group bought out IPowerWeb. IpowerWeb claimed it hosted 700,000 sites at the time. EIG ported all these domains to their Burlington, MA. servers, UNDER the guise of a control panel upgrade VDeck 3.0. Before I knew it my custom htaccess file was trashed by their port and they disabled a .htaccess feature "Options" requiring a re-engineering of the htaccess. Meanwhile of course the sites were still sort of running. EIG has bought up something like 30 hosting companies at least, thus my reference to how to avoid the PACMAN. Move to a new host, EIG buys them, it all starts over again! EIGs' server farm actually performs quite well, if they'd just keep their hands off the websites. They host 1.8 million sites now under 30 or so host names.
I reverse DNS'd one of my domains, finding about 100 other sites, virtually all of them had ads with pop under's smeared on their error pages, just like mine! Just think 1.8 million domains have had their error pages hijacked in the last 6 months. These are all small business sites, organization sites, sites that probably don't get a lot of maintenance, so their webmasters probably don't even know this has happened.
It's funny when I first started looking at this I was going to add custom error pages so I used the "IPower" .htaccess editor, and surprise, at the bottom of page in fine print was a "disable ads" check box! Confirming this change wasn't a hack.

If your going to choose a host look up EIG on the Better Business Bureau site. There is a list of all the aliases EIG is hosting under. But maybe you should choose an EIG host anyway, that way when EIG buys up more hosts you won't be forced to migrate!

[bumpski]

Getting back on topic.
I don't know about you but I do get quite a few misspelled in bound links. Links from scraper sites that don't know CASE is important!
So I was thinking more of Google purposely looking for sites serving pop up/under ads on error pages. The site visitor certainly would be getting a bad experience in this case, an error, followed by a hidden pop-under.
With the advent of Google's site preview this pop under detector technology would be a trivial enhancement.

So with my recent experience, to exaggerate, perhaps 1.8 million mom and pop sites could be hit be a Google penalty. That might shake up the rankings a little bit.

If Google is also crawling non-existent domains (another post) a lot more error pages with pop-ups would be served up. (I've completely disabled wildcard subdomains in DNS, again the KISS principle.)

[HuskyPup]

Better Business Bureau site link - Mods please delete if not allowed:

[bbb.org...]

[macas]

I don't know about you but I do get quite a few misspelled in bound links. Links from scraper sites that don't know CASE is important!
So I was thinking more of Google purposely looking for sites serving pop up/under ads on error pages. The site visitor certainly would be getting a bad experience in this case, an error, followed by a hidden pop-under.

About the same problem with my website .. I have more then sixty pages with that problem.
Now I getting worried about my rankings and affections on them.
Is there way to fix this or to resolve ? Anyone ?

+ Also I figured out that some of spam-auto-blogs are stealing my SE rankings and I discovered only because I changed few urls of my pages ... weird and confusing things are happening.

[TheMadScientist]

Yes, you can correct case using Mod_Rewrite, in fact jdMorgan made a very lengthy post with an example of how to do it, which I think can be found in the Apache Forum [webmasterworld.com] Library... Links to the Library sections are under the breadcrumb nav.

I definitely suggest having a look for it, and if it's not there, try your favorite SE or ask where it is, because I'm sure he'll know.

You can also correct basic misspellings in many cases, but that's a bit more involved and I'm not sure if there is anything posted about how to do it or not, but it can also be done.