User Testing: Does Your Site Need It?

Commentary by: Mike Osswald, Strategic Director

So you’ve spent a lot of time with your content - editing it for brevity - bulletizing key points. Your graphics are sharp and have small file sizes. Your server is fast. You've submitted your site to key search engines.

But how are your site's structure, navigation and overall usability? Do you really know? Unless you're using some kind of feedback service (such as BizRate, OpinionLab, or a Web-based survey), you probably don't get much positive or negative feedback from your site's visitors.

That's where user testing comes in. Unless you test with actual users, who haven't seen your site before, you don't really know how effective your site is at meeting your metrics. Even if you think your site is doing a great job, how much better could it be? While this means that anytime is a good time to conduct user testing, it's typically done before a redesign, and with newly-designed paper or Web-based prototypes.

The complete cycle of formal user testing requires planning of the appropriate users, appropriate questions and problems to solve, and appropriate data analysis. Testing is usually conducted in a private area, free of distractions. Video-taping and/or two-way mirrors are often used. There is a formal science to usability testing, for proper question generation and data analysis. See the Related Items section for more information.

Conducting Your Own User Test

That said, conducting your own “quick and dirty” usability test is a relatively simple process. First, you need to have an audience of typical users. Studies have shown that you really don't need more than 4 or 5 general users to point out the most obvious problems. You can use your friends and family, but they should be part of your target audience. Later on, as you refine your site, you can test other types of users (different audiences), and test specific areas.

As for the study, you need to have a list of problems or tasks for them to solve. The tasks shouldn't be too easy, and they should represent what your site is created for - finding detailed product info, evaluating criteria to make a purchase decision, making a purchase, finding a dealer, etc.

Let each user know they'll spend 45-60 minutes answering a number of questions. Let them know the purpose of the study is to find out what things work well, and what doesn't - be sure they know that they can't “fail”.

While testing these users, you will be watching out for problems in your site, such as:

  • Logical organization of content - how often do people get lost?
  • Understandable navigation - visually and verbally, are the navigation tools clear?
  • Usable navigation - does the navigation take you to major sections and key areas?
  • Page structure - do people read your headlines and subheads if they are an important part of content?
  • Click-depth - how many clicks does it take to get to important information?

You'll want to note everything the user says, and watch the screen carefully. You should note any delays and possibly mouse locations, as they will be indicative of sticking points and areas of confusion. After the test you'll want to review with each user why they did certain things when they got stuck.

Results can reveal major problems with structure, simple wording change suggestions, or even something as simple as the need to use bold to highlight important information. But, regardless of the good or bad results, you'll definitely find it to be an insightful experience.

Related Items and Online Resources: