The Human Side of Product Testing

The hardware has been tested and works reliably. The software has been tested and is robust, reliable and bug-free. Leading edge technologies have been deployed. The hardware and software are integrated well together and the functionality satisfies customer needs discovered in up-front market research. Cost and schedule goals are met. What’s wrong with this picture? Why is the product still vulnerable to failure in the market place?

Do you notice the lack of system thinking? What components need to work smoothly together for the product to perform well? You (probably) guessed it! The hardware, the software, AND THE PERSON USING THEM. A product with the slickest technology, the most reliable hardware, and the most sophisticated software just sits there without that other component of the system. A user must interact with the product to make it perform. And the interactions must be smooth for the system to perform well. A perceived or actual user-hostile interface will frustrate people, make them feel inadequate, and cause them to bad-mouth the product to friends and drive them to choose a competitor’s product the next time.

The hardware was tested, the software was tested and the software/hardware interaction was tested. Omitted from testing were the hardware/user and software/user interaction. This probably reflected the initial product requirements that specified the hardware and software in detail, but simply said the product should be user-friendly. There are several methodologies that can be employed during the development process to design usability into a product and to test for it. In this article I will briefly discuss one of the most widely used evaluation methods – usability testing.

Usability Testing

Usability testing is a hands-on approach to discovering the difficulties people have when trying to use the product. Design improvements are then implemented to remove those difficulties. Usability testing should ideally be done iteratively during the development process. That way each test is easy to do and takes less time because (1) only a segment of the user interface needs be tested and (2) only a small number of test participants need to be run. Just like with hardware and software testing, early usability testing identifies problems before they become so implanted in the product that they are difficult and costly to change. Early in development the testing can be done on a representation of the concept for the user interface. For instance, a “paper and pencil” usability test of a software user interface requires only sketches or printouts of proposed screens layouts. Another early method is Wizard of Oz which substitutes a person “behind the screen” for software that isn’t yet written. The person responds to inputs from the test participant. Later in the development process testing can be done on prototypes or engineering models.

Usability testing evaluates the user interface by focusing on user interactions and perceptions. The testing procedure is basically the same regardless of the testing platform.

  1. . Recruit participants who represent the type of people who will use the product when it is on the market
  2. Identify real tasks that exercise the interface you want to test
  3. Administer the tasks and observe the participant’s performance
  4. Encourage participant to “think aloud” while performing some tasks and solicit feedback after the tasks
  5. Analyze data to (1) find interactions that caused difficulties for most participants and (2) identify the causes of the difficulties
  6. Modify user interface design to remove causes of the difficulties

Usability testing can be thought of as a type of market research because it collects perceptions of the product. Much like focus groups, testing sessions are usually done in a room with a one-way mirror and are videotaped. However, there are important differences. Usability testing focuses specifically on the user interface rather than such aspects as market positioning and pricing. It is typically administered to one person at a time. Focus groups can be used, but only in conjunction with a usability test.

Usability testing has led to such improvements in how consumables are loaded into a machine, presentation of instructions, imbedded software logic, software applications, and web site user interfaces. It applies to products destined for the marketplace and for those intended for internal use. In other words, it will benefit any situation where it is important for users to be able to (1) quickly and easily learn to use a product and (2) use it productively when the learning period is over.

Stan Caplan