Test early, test frequently

 Introduction

We can not actually get inside the heads of users to find out if a navigation scheme works or if an interface is intuitive. We can, however, test.

As soon as a prototype or working system exists, tests should be conducted to check that the design works with the users of the system. Problems should be identified and corrective action taken as early in the design process as possible so as to minimise project risk.

 What and how to test

Use a representative group of actual users in the environment where the product is used. i.e. if they use the site or application at their desks, then test at their desks.

Test against original usability objectives

Do not ask the user for their own subjective views. Instead get them to complete tasks, and observe & measure the user against predetermined measurements.

Common 'general' measurements of the usability of a site or application are:

Effectiveness: How completely and accurately the user was able to complete the task, and what errors were encountered along the way.

Efficiency: How easliy the task was executed by the user.

Satisfaction: How much the user likes or enjoys using the site or application.

 Output from the tests

From these tests, we are looking to identify problems, and to gain a sufficiently detailed understanding of these problems to help us decide on corrective action.

The only output we need from the tests is an accurate and complete description of the problem. Exactly what form that fix should take can be debated with other members of the project team after the tests.