Redirecting the user is when it does become absurd. Because now your web server is sending a redirect response, which not only results in the browser having to make an additional request following the server response’s ‘Location’ header, there’s still no indication that this was a 404, which is clearly is. It would be one of the 3xx status codes. This does not preserve web server semantics, it ends up breaking it. Redirects are semantically used to direct people to the new location of a previous resource on the server.

You can serve a “not found” page to your users and still have your web server send the 404 status header, making the response a 404. Some browsers (like Chrome) will still choose to display their own 404 pages while others will display what’s been provided by the web server’s response body.

In short, this is bad advice from the perspective of preserving web server semantics.

P.S. This is an awesome way to break SEO. While it’s unlikely it’ll have a huge impact but by doing things right you can improve your SEO little by little. Please take consider and take some time to read the HTTP 1.0 and 1.1 specs.

-Anon

Sorry to ask in comments, but i havn’t been able to find any information about this anywhere.

Im currently doing it in PHP, which basically checks that the requested path isnt /index.php, but isnt it more reliable to do in heaccess?

Just curious if (and how) that affects our bounce % statistically.

First, let me just start off by saying that this post was meant to inform/remind beginner eCommerce Webmasters that there are other options beyond creating a custom 404 page and that once they are aware of both, it’s up to them to decide which method would work best within their business model. I believe I mentioned that for users who already had a custom 404, that the redirect would probably not be the best method for them since much of the work is already done. But for merchants without anything in place, that this would be one quick and easy method for keeping customers who encounter a 404 error on the site.

Many of the things you mentioned about the custom 404 are present on most website index pages, minus the url location and matching anchor related to just that page. eCommerce index pages today look more like portal pages than home pages. They usually have more than 100 outgoing links, they are usually linking to the most important sections, there are a variety of keywords describing the entire product line and there is almost always a search function with appropriate navigation. The only thing missing would be the message informing the customer.

I think as Internet professionals we sometimes forget how much more savvy we are at navigating the Net than the average Web user. An average user, who is shopping online, doesn’t know what a 404 page is or what it means. All they are interested in is finding what it is they are looking for. I have even seen a few rare occasions where custom 404′s have the highest bounce rate out of any other pages. Both of them used custom 404′s that had all of the elements above in place. When they switched the custom 404 to an index page redirect, they noticed the overall bounce rate decreased and overall page views increased (can’t be absolutely proven that it was completely due to the switch though).

I believe there’s no one size fits all 404 page solution. I think depending on the customers you target, your program and website goals that either one is better than nothing at all.

I have seen some pretty awesome things being done lately with 301′s and 404 page includes, and have been experimenting with a few ideas along the lines of what you said about performing internal searches based on the original page file name or title. Can’t wait to discover some new strategies. I liked your idea about live help.

I guess maybe the title of this article was a little misleading. The word ‘better’ should probably not have been used. Perhaps ‘alternate’ would have been more appropriate.

Thanks for the comment Jason. If you decide to blog about this I will definitely join in the discussion.

Ideally, the URL structure should give an indication of intent. i.e. http://www.domain.com/categoryX/missing-product.html

A site can offer up the top categoryX results (sort by highest rated where possible), or do an internal search on the missing-product.html file automatically and do a redirect (301).

With product churn this is common, so really advanced stores can keep record of product URLs and automatically forward to replacement product, or cross-sold product (but always tell the shopper they have been automatically directed to this product because…).

So I do not like sending a user to the homepage upon 404 because there is no anchor. If the resources are available I’d have live chat operator automatically help users who hit a 404 too. Hmm, maybe I should blog this topic too.