This week it occurred to me that one of our analytics posts, Tracking Error 404 Pages and Broken Links in Google Analytics, actually had some pretty hefty SEO implications.
Error 404 pages are something that every webmaster has to deal with. Best practices regarding them vary, as most things do, depending on the field you work in. Redirect? Make careful custom 404 pages that invite users to keep looking? Contact webmasters of linking sites and ask them to change their links? What to do.
The middle ground approach seems the best, with different actions depending on the page, who is linking, and what you think a user is trying to reach. However, interpreting and managing that kind of information is a full time job in and of itself. Fortunately tracking your 404 error pages in Google Analytics gives a quick and easy way to prioritize your action and pick off some of the low hanging fruit.
The problem with 404’s
No one likes a 404. A default 404 error provides a poor user experience, and gives few opportunities to redirect visitors to an appropriate page.
bit of a brick wall, no?
Custom 404’s, on the other hand, can serve to guide users back to the content they want to find. Check out Jeff Atwoods advice over at codinghorrors for some good advice on building 404 pages, though I notice he never mentions my favourite 404 addition: search boxes. This provides a valuable way for users to find content that they might otherwise just skip trying to get at.
The big debate with 404’s is whether we should be redirecting links. The SEO implications are obvious: people linking to non existent pages are a potential SEO source that you aren’t tapping. A 301 redirects that “link juice” (or “pagerank” or whatever) back to pages on your site. The problem is that blanket 301 redirects cause a bad user experience (“I went to /favourites.html and got the homepage? What the heck is this?”) and gives some potentially poor signals to Google. (though whether Google acts on those… who knows)
Matt Cutts advice is classically Cuttsian: using Google Webmaster tools, see who linked to your 404 pages, and ask them to change their links. Bam, one email, “free” links. If only life were that simple.
For those of you that have big sites with heavy folder structure, and potential for tons of 404 links (say, a website that uses Canadian/British spelling for its page names) you’re not going to want to contact every one. Sure, you could prioritize based on which pages are getting the most/most trusted links, but are those pages sending traffic? How much? What’s happening with the traffic?
As SEOMoz suggests, the solution may be an intermediary. Let some users hit a 404, 301 some users, and contact some webmasters. Which do you do for which? Ask Google Analytics.
Analyzing your 404’s
If you’ve got Google Analytics installed on your site—and really, given that it’s free you have no excuse not to—add the GATC to your error pages (bonus: adds to your 512b requirement for a 404 to show in IE and Chrome). Be sure to title your 404 pages appropriately, so that you can create filters to see your 404 pages. Then from Google Analytics drop a custom filter such as:
Now you can pull out as much data about your error 404 pages as you like, including where they are linking from, keywords & Search Queries, how they got to the site (direct, etc.), mediums, sources, etc.
I like sources, myself:
You could also compile these into a custom report to provide any sort of information you may find useful. Perhaps visits, bounce rate (how many people are being driven away by your 404 pages) and visitor type. Then for pages of interest click through to see where they’re coming form and where they are going to.
This nicely augments your Google Webmaster Tools information.
Using your 404 data
The uses for this data are endless. Some simple ones include:
- Are you seeing that a link to a missing page sending a lot of visitors? Maybe you don’t want to 301 that page, but rather contact the webmaster and have them change their link to point to the proper page.
- Have moved content that’s hasn’t been registered as a 404 in Google (so there’s still search query data attached to that visitor)? 301 that page to match what they’re searching for.
- Have a bunch of users going to a URL that doesn’t exist, but is a close match to one that does? 301 it to that URL. I have to guiltily admit that for some reason we still haven’t 301ed vkistudios.com/blog to blog.vkistudios.com.
- Are people visiting a page that doesn’t even resemble anything on your site (this happens more frequently than you might think)? Use a 404 error page with some top results and a nice big search bar.
- Is a major site with high trust linking to a page that doesn’t exist? Ask them to relink, and if they don’t respond: 301 it.
- Are a bunch of decent quality sites linking to a page that doesn’t exist, but not sending a whole lot of visitors? 301 it, saving you the effort of contacting every page.
There’s loads more uses for this, and I am sure some of the more SEO savvy or analytics savvy readers could think up some really great ones. (and could even leave them in the comments!)
