Bringing it together - The Usable Site Part II

Usability testing

It is a good idea, however small your site, to do some sort of usability testing. 

No site, however well designed, is totally without usability problems form the outset. It can be painful to have some of the design elements you really like and have worked hard to achieve turn out to be a problem for users, but it is far better to discover this at an early stage than later. 

Setting up a small usability study yourself is not complicated and can be extremely beneficial to your end result.

Choosing Testers

Surprisingly only a very small number of people are required for usability testing, even in professional corporate testing situations it is reckoned that just 5 to 7 testers is usually sufficient. Recruit some friends or family members but try to get a good spread of experience level.

Setting the Tests

Asking people to look at the site and find problems is a poor approach, the question needs to be a bit more sophisticated than 'what do you think?'! Far better is to create a number of user tasks. A user task is a description of what some user may want to do at your site. You should come up with a task for each of the groups of potential users you identified previously. 

On an e-commerce site a task might be to buy a particular item, perhaps with specific additional features, and send it to a predetermined address. On information based sites a task might be to use the site to answer a question. Vary the level of difficulty of the tasks, as you perceive them. At least one task should be to ask people to find something or do something that might be possible, but is not, on your site. Tell them that one of the tasks is incapable of being completed at your site but not which one. 

While setting these tests has some value if people carry them out in their own homes and report their experiences to you, it is far more valuable if you can observe the testers on your site.

Observing the Test

The first rule of observing is to stay quiet! Resist the temptation to guide or suggest options, in fact if you feel the need to do so recognise that you have identified a usability issue. Make notes throughout the test watching in particular for the points below as each task is carried out.

Look out for the following:

• How quickly do they choose the appropriate navigation option? If they spend a long time considering their initial choice are their too many items or are the items badly named? Do they choose a wrong option first before finally getting the right one?
• How many screens do they go through to reach their destination? Is there a way you could have taken them there more directly, perhaps by providing shortcuts or alternative navigation?
• If the information they want is not available on the site how long did it take them to realize this? Did they wrongly identify an item as being the unavailable one when in fact the information is available?
• Are they straining to read text? Does it seem to take them a long time to find the particular information when they arrive at the correct page?
• Do they get lost or disorientated at any point? Are they able to move between the various questions without difficulty?

Measuring the Results

The simplest way to measure results is to award one point for successfully completing a task and zero for failure. Thus the result for each task will be the number of testers who managed to complete successfully. 

A variation is to ask the users to grade the difficulty of completion from 1 to 5 and you assess your opinion of their progress on the same scale. When you later add your score and a users score for each task you have a measure in points out of 10 for each task. If you repeat the test later you will have a base result by which to measure improvement.

Acting on the Results

It is pointless to test and then ignore the findings, though a surprising number of people do this. If it turns out that your navigation is confusing it is difficult to throw aside the many hours you spent making lovely gifs for your buttons and redo the lot again. The thing to bear in mind though is that the site is not really for you, it is for your users and their needs must be placed ahead of your preferences.

<< Part I

Katherine Nolan
OutFront.net
A Microsoft FrontPage Learning Community