Suomeksi | In English
Interacta Design Oy
GUIDe process model
Services
Testing usability
Training
Publications and articles
Customers
Contact information

Testing usability

At first, it is worth assessing whether it is more beneficial to start improving usability by testing or redesigning the existing user interface.

If you know that the current version contains many usability problems, or that you have never been able to focus on designing the user interface, creating a new user interface is likely to produce better results than testing the current one (see Designing user interfaces). Similarly, quickly written, low-quality source code is not worth testing and patching again and again, but instead it is sometimes better to rewrite it to create a better structure and organization.

Testing brings forth problems in the user interface

Complex navigation path
Image: A complex navigation path of
a test case.

When we test a system with test cases that are based on real use situations, we can systematically find out the problems in the user interface that prevent the users from fully capitalizing on the systemís services. We report the problems found by testing with illustrative screenshots.

Once the user interface problems have been discovered, the most demanding task is still left, i.e. designing the improvements to the user interface. This is the most essential focus area of our company. Our test reports always include suggestions for improvement, which have been tested by usage simulations and are reported with screenshots.

Simulation-based testing without users

Simulation-based expert review depends on an experienced user interface designer who is able to predict problems that the users stumble on by simulating use sequences many times over. Simulation-based testing discovers especially effectively problems of long-term use, which the users would encounter having used the application for a few days or weeks. A typical example of a problem of long-term use is an arduous, error-prone and long sequence of steps to complete a simple task (e.g. 70 steps to handle a damage report). Usually, such sequences can be easily shortened (e.g. down to 15 steps), or even partially automated. In addition, an expert review is good at finding problems of missing functionality and unsuitable organization of data.

A sample excerpt from an expert review report of Ical calendar application, created for training [pdf 315 kt in Finnish].

Testing with the help of users

Especially the usability problems related to first times of use (learnability and intuitivity) can be found out by conducting usability tests. In a usability test, a test user carries out test tasks and tries to use the target system as he best sees fit. Usability problems are figured out by analyzing the actions and verbal comments of typically 5-6 test users.

Usability walkthroughs are easier and more inexpensive to organize than usability tests and are especially suitable for early-development-stage GUI drafts and prototypes. Instead of watching the actions of individual users, the use of the system is studied in a supervised meeting with 2-5 users.

When planning for user testing, one should note that the results are heavily weighed on the experiences of the first few times of use, and problems that occur during long-term use may remain uncovered. Such problems can be discovered by conducting user observations, where the use of an existing system is observed in the real environment for a few hours. User observations are especially suitable also for evaluating the user interface of mobile devices, for example.

Test cases and the choice of testing method

All the testing methods that we use are strongly based on realistic use sequences. The test cases correspond with the real-life situations that the users encounter daily when operating the system. In many cases it is worthwhile to conduct a couple of user interviews to find out these concrete use situations before the testing is started. Unrealistic test cases can easily produce erroneous results.

We often recommend having us first conduct a simulation-based expert review, because it quickly reveals the worst shortcomings, which can the be corrected before possible user tests. The selection of the most suitable testing method is heavily influenced by the purpose of the application, the stage of development and many other factors. Ask for a recommendation of the most suitable testing method for your system by either calling or emailing us.