How should a website handle errors? What kind of information and messages should be presented to the user? And will a one-size fits all approach do the job?
Our own 404 page is pretty lame, so we’ve been thinking about the various ways in which we can improve things, as part of our year-long project to provide you with a glorious new E-consultancy website (due in early-2009).
I have also analysed the 404 pages of 10 of the most popular news sites, to see how other publishers are faring.
There are theoretically three different kinds of error pages, and each one of the above should have a custom page, designed in accordance with the type of error. They are as follows:
1. You screwed up.
Doh! It is important to let users know that it might be their fault. Ask them to check the URL. Most of the 10 websites I looked at hated capitalisation in URLs, and couldn’t handle it (something I wasn’t aware of previously). That’s the kind of mistake that somebody could easily make if typing in a URL. Like ours, some of the sites I looked at provided few clues. The FT simply claimed that there was a ‘server error’. Obvious links back to key site sections and a search box should be mandatory.
2. We screwed up.
Oops! Something broke. Many 404 pages contain a ‘report a technical fault’ link, but why not write a custom page when the site breaks? Hold your hands up. Again, prominent links / navigation and a search box should be included to help the unfortunate user stay onsite.
3. We’re fixing the screw up.
Hang on a minute please. Again, if you know that the site is likely to be offline for 45 seconds while the server restarts or the database rebuilds then why not advise accordingly? There’s no point placing a search box or any web links because you know that the website won’t work. This page might simply state ‘we’re conducting maintenance, come back very soon’ message. Or it might include a picture of a man in a builder’s hardhat.
So how do news organisations handle 404 pages?
I selected, at random, 10 news organisations from the US and UK, to see how they manage errors. Note that these are standard 404 pages and since I focused on URL errors I can’t say whether they have the two other types of page for site errors or site maintenance. It was my fault, not that the FT blamed me!
Trends? The US sites tend to feature a bit more navigation and a search box, which is best practice by my reckoning. UK sites seem to favour sparser 404 pages… sometimes a bit too sparse for my liking.
For me, the key things to get right are clear messaging, browsability, search, and error reporting.
I have awarded each site with a score out of 10, just for the hell of it. Our current 404 page is probably a 3/10 at best – poor messaging, but redeemed slightly by the sheer amount of navigation and search functionality. We will improve this in the coming months.
Ok, from worst to best…
Sky News: 2/10
A blank page, a pithy message about a ‘server error’ (rather than a user error), and a Sky logo that doesn’t bother to link to the homepage. The saving grace here is the five text links to various key sections of the Sky website. Nothing else going on at all.
Removes almost all branding and navigation, to display a 404 message on a monochromatic page. Click on the Telegraph logo and it will take you back to the homepage. The messaging is pretty good, but that’s about it. There are no text links. There is no search tool. The browsability factor isn’t great.
Additionally, the page includes a banner ad at the foot of the message, which seems to me to entirely miss the point. The ad for ‘Lakeland’ took me to another 404 page. I wouldn’t want my ads shown on a 404 page.
The Times: 3/10
Removes all branding apart from a TIMESONLINE logo that doesn’t link to the homepage, for reasons unknown. There is however a text link to the homepage, and a ‘link’ to ‘Register a technical problem’, which is actually an unlabelled email address. There are no other links, no search tool, no
But there is a picture of a green box with 404 written on it, with what appears to be a yellow parrot perched on top, and, worse still, an advert for a promotion hosted elsewhere on The Times website.
I also stumbled upon another near-blank page that simply stated ‘Site is currently unavailable. Please come back later’. And that was after I made a URL error, when the site was definitely available.
Removes almost all branding and navigation, to display a 404 message on a white page. The message includes three in-page links to the News and Sport homepages, as well a directory of all BBC sites. But the logos don’t link, and there is no search tool. A bit limited, but the messaging works ok.
The Guardian: 6/10
Removes all branding and navigation apart from a guardian.co.uk logo that does not link to the website. An easy-to-digest list of text links are presented on the right side – these take you to sections of the site. Straightforward enough, but leaving the top navigation in place would have had the same effect. There is a contact link ‘if you require further assistance’. Disappointingly, from a site that gets it so right so often, there’s no search tool.
The FT: 6/10 (despite the weird messaging)
An interesting approach. The FT’s 404 simply blames the server, rather than a user error. Underneath the heading ‘SERVER LOAD’ it says: “Our server was unable to complete your request due to high volume. Please try again by clicking your browser’s reload button. If you receive this message again, wait a few minutes before attempting to access the page again. We apologise for any inconvenience.” Hey, maybe it wasn’t the server’s fault!
It could be more transparent. Educate users, if you need to. The other aspects of the page are good: a search box, links to the site map and homepage, contact us, and logo that links back to the homepage (as well as no less than THREE text links that do the same thing).
The Sun: 7/10
Retains some navigation at the top of the page with links to various sections of The Sun website, including a logo that links back to the homepage. There’s a search box, a text link to the homepage, an email us link to report a technical problem, and a bunch of links in the footer. The messaging is clear. Pretty, pretty good.
The Daily Mail: 7/10
This is a decent approach, with two exceptions: The Mail simply pulls through the header and footer from the previous page. So the top navigation is left intact, apart from the search box that lives just underneath it, but the browsability here is very good. The navigation highlighting is preserved, so if you’re in the TV section when the error happens, the navigation on the 404 page continues to highlight ‘TV’. The bad news here is that there’s a picture of Lily Allen in the header.
In fact it is twice as good as it could be because the top navigation is replicated underneath the error message (which states a ‘technical error’ rather than a possible user issue). It was actually my fault that I mistyped a URL.
There’s also a text link back to the homepage, and a bunch of links in the footer. Probably the most links on any of the 404 pages I’ve looked at today.
But there are also three ads displayed, which isn’t entirely necessary. That was one of my exceptions. The other one was seeing Lily Allen, though that hasn’t affected the scoring.
The Washington Post: 7/10
Retains the top navigation to allow easy browsing back to a section of your choice. Also includes a prominent search box in the middle of the page. There’s a ‘report this link’ form. Not too much in the way of an explanation, but it’s a winner in terms of functionality.
New York Times: 8/10
It also includes a ‘Go to a Section’ dropdown, which will keep many users on-site. I like that a lot. The messaging is ok. Overall the page is simple and effective, focusing as it does on search and browsability.
Chris Lake is editor in chief at E-consultancy and can be found Twittering here.