« We're #1! We're #1! | Main | Google search operators »
Custom 404 Error Pages and Deceptive Error Codes
September 27, 2005
Last week, a client asked me to look into a strange search result on Yahoo. For this particular search phrase, the page listed in the Yahoo SERP had long ago been deleted. In fact, the Yahoo descriptive snippet for this page gave the typical "error" description from the client's custom 404 page:
"The page you requested may have moved or no longer exists...," etc. etc.
404 error pages sticking around in the engines' indexes is nothing new. Many clients wait 3-6 months or more for their 404 pages to drop out of Google and Yahoo. But this was different. I ran the client's URL through a header checker, and sure enough, the problem was not a stale index. The problem was a deceptive http header code.
Many webmasters believe that creating 404 error pages automatically ensures that the server automatically passes the 404 error code along with it. This isn't always the case. Here are a few examples.
Adidas error pages give the http 200 code. The 200 code means that the intended page was found, which is certainly not the case in a URL like www.adidas.com/zzzzzzzz. The error page does a meta refresh after 10 seconds back to the Adidas home page, so the user eventually arrives at navigable content. But it really should send the 404 code instead to ensure that the legacy URL gets flushed from the engine's index.
The British National Space Centre uses a 302 redirect to take the user to a page containing error copy, but that page gives the 200 error code. (Try this fictitious page to see what I mean.) So for any BNSC pages lingering in search engine indexes, the content of the first page is used to create the SERP description, and the second page is the user's ultimate destination. The problem is that engines frequently believe that both pages still belong in the index, when in reality neither does.
In the health checklist of SEO, this issue isn't equivalent to a broken leg or clogged arteries. It's more like an old football injury that flares up at inconvenient times. The most likely results of a deceptive header code on your error pages are search query results that return error copy in the SERP description, and possibly an artifically inflated index count. Neither one will kill you, but it's something you shouldn't ignore.
I've found this phenomenon to be most common on Microsoft IIS platforms, usually with .asp or .aspx pages. This doesn't mean that other platforms don't also cause the error - only that I haven't seen any non-MS instances.
Make sure that your error pages produce the proper 404 header code. To do this, create a nonsense url, such as www.mycompany.com/fake-page-name.zzz - in other words, a page that you know does not exist. Insert that URL into a header checker, such as the one at SEO Consultants or Rex Swain's HTTP Viewer. (Rex's site has been on my bookmarks list for several years now, and I use it about 10 times a day.)
If your fake URL gives any header code except a 404, it's time to roll up your sleeves and fix it. Here's a long thread about possible fixes on IIS, and WebmasterWorld has a little information about the same thing happening on an Apache server.
see all posts by Erik Dafforn
posted by Erik Dafforn at September 27, 2005 04:30 PM
Intrapromote: [ Case studies | SEO services | Bios ]
Trackback Pings
To TrackBack this entry, use the following URL:
http://seoblog.intrapromote.com/mt-tb.cgi/67