Don't Forget to Test!

In writing today, I was reminded of issues with respect to website usability. I seem to have gone through a dry spell with nothing new to add to this blog or a lack of time in which to do it. Hopefully, this entry signifies the end of the problem.

Students had a great deal of difficulty navigating the site of an on-line school. The homepage included directions and there was a tutorial, but it didn't seem to help. Were they really being used? The message boards didn't seem to be read. Students did respond to emails. Was terminology the problem? What was the problem with the navigation?

Interaction using Blackboard created a real-time interface, but it required that the instructor's and students' schedules match up. Several instructors worked with regular schools during the day and were only available in the evening. Some students were taking the course at school and didn't have computers available in the evening. Time zones were an issue.

The site was very graphic intense for those students on dial-up connections with older computers. The slow loads were a problem. Rural areas didn't have the option of broadband.

When talking to a member of the company, I was told that the programs were new and this was the first that they were being used. I have always assumed that such programs were tested with students of similar backgrounds and connections, but the problems make me wonder.

Usability requires that we address issues of:

• appropriate reading level


• connectivity


• computer skill level


• text versions of interfaces


• access or limitation thereof


• human-computer interaction (HCI)

We test to find out what we have overlooked as well as what doesn't work. The latter isn't just due to logic or code errors. We know what the sites are supposed to do and where to find things. It is very important to be sure that our focus groups and usability testing be run with an appropriate sample. Videotaping the individuals provides a back-up to catch expressions and actions that might provide clues to problems. Once the program and project were ready for release, a pilot run would have prevented difficulties for a large numbers of frustrated and disgruntled clients.


copyright © November 2004 K.E. Lubrecht
All rights reserved