Faulty websites cost cash

Are you losing money?

[georged]

According to British 'newspaper' The Mirror, faulty websites (presumably including 404s etc.), which surfers cannot log onto, cost british business £128 million last year. They say that's according to an ISP called Hostmark. And that's all the info they give.

[rogerd]

That sounds like a wild guess if I ever heard one! At least it would be hard to disprove...

[Brett_Tabke]

I would believe it. How often do you turn around and go back when you hit a 404 while looking for specific content?

[knighty]

Makes sense actually just think how many sites there are that crash/freeze as you press 'confirm vast amount to be spent'

I remember trying to buy a DVD but the thing froze and i never went back, on that one fault they lost my $20.

[mivox]

A lot of times, it's the 404 page itself that decides it for me... if the 404 page is totally generic, or has no useful information, I leave. I've hit 404s with links to major areas of the website though, and I'll often use one of those links to give it another go.

For my employer's site, I put the basic navigation bar and the site map on our 404 page [absak.com]... so you really can get to everywhere on the site from there. When I look at our logs for 404s, the individual visitor paths most often show someone entering a faulty URL, getting the 404, and then going to the correct page (or another page on the site), rather than leaving. Seems to be doing the trick.

[Bjorn]

Except for the figure which cant really be validated I agree fully. The problem as I see it is finding the problem areas of a site. Personally I tend to get a bit blind to the things I work with all the time.

Also it seems that it is more fun to build something new than it is to fix something that doesent work.
Well, IMHO anyway.

[2_much]

Hi Bjorn, welcome to WmW!

I find myself having the same exact problem. It's hard to step aside from my paradigms and schemas and to think and carry out tasks the way other users would.

Jakob Nielsen has written some articles on usability testing and ways of finding out how other users react to your sites depending on their level of experiences. When I find one I'll post it.

[mivox]

> has written some articles

LOL... *some* articles? :) I think he wrote ALL the articles on those topics...

And Mr. Nielsen's expertise aside, the best thing you can do for your own sites is to watch your logs like a hawk (IMHO). If a lot of your visitors seems to leave in a certain section of your site, go look at it and ask yourself why. (Unless it's your links page, or the last page of the checkout routine in your shopping cart... ;) )

If you change something on your site and see the average # of pageviews per visitor drop from 5-6 to 1-2... you goofed. Go back and undo the change.

And take the time to build custom error pages for your site... at LEAST build a custom 404 page. Give a link to contact the webmaster, and links to other major areas of your site... make it easy for visitors to get back on track (or email you to complain) when they hit a 404.

[knighty]

how do you build a custom 404?

[mivox]

Build your 404 page just as you would any other html document, save it as 404.html, upload it to your web directory, and add the following line to your .htaccess file:

ErrorDocument 404 /404.html

[Marn]

Mivox, editing .htaccess does not work on Win2K or NT servers, you have to edit IIS under Custom errors to point to the URL of the error page.

Below is a link to a great page dedicated to 404 pages, and he has even more detail on how to set it up on various flavours of webservers.

[plinko.net...]

[rcjordan]

I used javascript to write the referrer into a hidden field and ask the user to submit a "trying to find" comment form from the 404 page. I get some type-in errors where they just made up a logical url and tried it (very informative on finding interest hotspots I may have overlooked). Mostly, I just get a comment giving a key word phrase or a topic they thought I might have on site. But, if I see a referrer in the hidden field, I know there's a problem out there. You can't always fix it since they are often old pages still in SERPs or links on sites that are all but abandoned or poorly maintained.

[mivox]

editing .htaccess does not work on Win2K or NT servers

Very true... thanks for filling in the blanks for me. I've never hosted on a Windows box, so if it's not *nix, I probably don't know how it works.