In-house Usability Testing Essentials

Basic things you'll need to start

Some of the essentials of in-house usability testing are as follows:

The test waiver:

Make sure your test subjects sign a waiver. You need approval to take notes and certainly to videotape or photograph them. If you do not record their interactions you will not be able to review the usability test subject responses with others. Your understanding of problems, etc., will get watered down by opinion, lack of adequate recorded data, problems of subjectivity, etc.

Preliminary questions:

Collect demographic data on your users and be sure that they closely match those for users of your site both presently and in the future.

Task based questions:

This is the heart of a good usability test and coincidentally also the place you are most likely to go wrong: you are most likely to get them to do tasks that you want them to perform. It is extremely difficult for people behind an interaction to create really useful task questions that are not leading or simply irrelevant to the user.

If you base your task-based questions on others or on other interactions, it will not identify what users find uniquely useful or unusable about your interaction. We ask you what your user is supposed to do on our site and we ask you what your goal is for them. In creating task-based questions, we might also use our own interaction with the test subjects to modify tasks; this may take place over test iterations or the issue may arise immediately. Again, the objectivity of professional usability test presenters is critical to completing the goals of the test.

Here's a cheat sheet of some sample task-based questions, courtesy of the University of Texas.

Test your usability test!

If you`ve developed a full usability test, start to "test the test" by taking a dry run yourself. Proofread questions and go though the usability test, testing out whether the question, the tasks assigned and all other elements really make sense. Make sure the site is working and all the files are in place so that the test will not fail unnecessarily, requiring you to ask people to come back or find new test subjects. Make sure that the test will not run into any obvious dead ends and that there is a sure path from start to finish that a user could follow.

You need to also practice giving the test. This will actually take a bit of acting, since you will need to present objectivity and disinterest in the test results. Have a co-worker or a friend as a test subject, complete all the paperwork, listen to your instructions and even run through the test to "test the testing" a second time. Try not to let familiarity with each other skew your assessment of the test and your ability to run a full usability test with real subjects. Practice digging for answers, and ask your test tester to be brutally honest. All involved need to simulate disinterest.

Supply checklist - make sure you have everything you need to conduct each test. Your checklist might look something like this:

• Test Waiver
• Entrance Questions
• Task Based Questions
• Exit Questions
• Computers with a variety of internet connection types along with all required plug-ins
• The URL of the working web site
• Gifts of appreciation (e.g. gift certificate)
• Pen(s)
• Other tools and implements as required
• All appropriate recording help (video recording, personal attendants, analytics software, etc.)

Ideally you should test with different connections, different speed CPUs and different operating systems (e.g. Windows vs. Mac).

In describing all this, we realize we sound prohibitive. From our perspective, we are attempting to be nothing less than professionally cautionary. If you want to launch a failsafe product, test it properly. The way to assure proper usability test results is to conduct them using usability testing professionals.

If you're ready to proceed, though, it's time to conduct the usability test or go back to in-house testing cautions (page one of our amateur usability testing mini-guide).