By the time users get to your error page, you’re already struggling about goodwill.
It doesn’t matter whether visitors get there via some old page they bookmarked, a referral from another site to a page that no longer exists, or some internal issue on your site. What matters is that now, your site’s user is upset. If you think dealing with a neutral visitor with the attention of a lit match is difficult, just imagine how much harder it is to deal with that same user now that he or she’s annoyed.
That’s the bad news.
The good news is that error pages have been around for a while now, and while we’re learning about new techniques to improve them all the time, there are some solid practices that every site should follow.
Know your “recovery” rate
Just about every clickstream tool there is—from Omniture, to WebTrends, to Google Analytics—has a measure for the exit rate. Here’s the thing: on almost any page on your site, the exit rate is useless.
The last page users view before leaving the site tells you nothing. A person who viewed the home page and left on a product page may or may not have gotten what he or she needed.
The exit percentage is pretty useless UNLESS it’s for one of these two areas:
- A step in the checkout process, or similar linear funnels
- The web site’s 404 error page
For checkout processes, the exit rate determines whether you have a leaky bucket, and you can fix the pages with the highest exit rates. For 404 error pages, the exit rate determines the percentage of people who see the error and never recover.
If you customize your 404 page and that number goes down significantly, you’ll know you’re on to something.
Be more helpful than creative
If you’re Coke, you can get away with error pages like this:
If a site is not very task-oriented, you can get away with cute. Just about everyone else who is not Coke should stay away from this.
Everyone else needs to offer more. Providing more is not that complicated: a good error page is easy to understand, accurate, and helpful.
In the early days of the web, error pages were these strange messages that very few people understood—nobody knew how to design them better, so everybody just kind of accepted it. But since then, the default 404 page become really outdated. One of the reasons is that it’s not hard to design pages like this:
Note the simplicity—the options stand out on the page, and you can either run a search, go back to the home page, or report the issue. It also helps that it was written for humans, and it’s not a type 4 error or some other esoteric message. What’s more, it genuinely points the user in the right direction. It also doesn’t hurt that eBay has a pretty solid search function right in front of the user when things go wrong. Search is very useful as a catch-all.
It helps to know their tasks
Just as eBay users are likely to use search, people traveling and using hotel sites are likely to, well, look for hotels. So, which of these pages is more likely to be helpful?
For web sites that are strongly associated to one type of task, the hotels.com model wins, hands down. The 404 error page allows the user to recover right then and there. And it’s not like booking.com’s error page is bad, either. It’s easy to follow and it does provide some options, but you have to get out of there to perform key tasks.
The example above is for hotels, but this applies for any commerce-based task on the web: if you know what your user’s tasks are likely to be, put those on your error page. You’ll help a lot of them.
Put a connected computer in front of a person, and errors are inevitable, even for sites that are really usable. The worst thing you can do for those people is say, “404 – File or directory not found,” and then not leave them any options.
That’s still what a considerable number of web sites do. Don’t be one of them. You need to customize your 404 page, base the content on known tasks, and check if your users can recover more.
It doesn’t take a lot of work to help your users precisely at the point when they need the most help.
This article originally appeared in Tim’s Monetate column November 18, 2014