Posted in: Archive
Broken links are important to resolve. They are the one area of your site that can damage your brand and leave your visitors with a bitter taste in their mouth.
Here’s what Google’s current ‘can’t find your page’ currently looks like. Many of you will already be familiar with it. You’ll see this when they remove a page or service, or screw up a link somewhere across Google properties and websites.
To me, for a company that specializes in finding needles in the proverbial web haystack – the traditional 404 with a friendly broken robot image frankly isn’t good enough. It is a cop out. It’s the same one page of useless you see thousands of other websites, its the shoulder shrug of responsibility passing from Google to the user. Its the ‘Whoops, my bad, we broke something. Tough’.
It’s the breaking of brand promise.
After all –That’s all we know isn’t quite the truth.
Here’s what Google actually know when you land on the friendly robot from another page:
1) They know what page you visited from to find the broken link on their site. (via the Referrer)
2) They know the keyword context of the page that you were on previously.
3) They know the link text of the link you clicked on to find the broken link (from previous crawler data).
4) They know the context of the page which used to be there. (again from previous crawls)
It would be trivial to deliver another search result as part of the 404 page based on this information, returning the user to a tailored Google search result that just knows where you were, and the context of what you were looking for based heavily around the link text and the context of the previous page.
Failing the more technical ‘Prediction of user intent’ approach, it could improved further still, with the simple inclusion of a search box returning the user to a Google search. After all, isn’t that what the majority of users do when they can’t find something anyway?
Granted this is an engineering task that isn’t going to be priority number one for them, but right now, this page is embarrassingly poor, and wouldn’t be that much of a technological leap to resolve. Indeed, many of the points they highlight themselves as necessary on a 404 page, clearly don’t apply to them.