I was recently asked to review Parimala Hariprasad’s CAST 2014 talk, and what a great talk it would have been; shame I missed it! The review prompted me to remember a technique that I’ve used in the past which I call Error Elimination Testing. I also thought it may be a good idea to finally write about it!
From past experience, it’s not uncommon for people to test for error messages. They invoke the message to make sure it appears, is readable/understandable, says what it’s suppose to say, etc, etc. But how many people think a little bit further, and test whether it is actually necessary, or think about possible changes to design elements which would mean there was no need for an error message in the first place?
I give you… Error Elimination Testing – Testing in order to eliminate the need for error messages.
One of the most common instances is the sign up/register page…
Seems simple enough, so let’s invoke an error message…
Fair call. 6 characters will be a bit safer than the 1 that I tried! However, if I’d used my dog’s name, which is 5 characters in length and far more likely than someone trying to enter 1, I would have got the same error.
So how much of a design problem would it be to have the text “Your password needs to be at least 6 characters long” permanently displayed on the page? I’m not a web designer, but I can’t imagine it being a big deal. A small line of text to save the frustration of being presented with an error message, and having to enter another password… works for me.
The above example is a simple one, but what a difference it could make to the overall user experience. When you combine issues such as these in the sign up process with other potential issues across the site, they can add up to a level of *insert negative emotion here* that detracts significantly from the user’s experience.
Once you do this on a few occasions you’ll find yourself thinking about it every time you see an error message; “What could we change so that we don’t need the error here?”
There will be times when you can’t change anything, and that’s cool, but at least you’ve taken your thought process one step further.