Usability Articles - Human Computer Interaction and Interface Design

Tuesday, 26 February 2008

Print

Usability Testing

Usability testing is one of the most important parts of the system's development life-cycle, but it is often the most overlooked. The problem is staring us in the face - developers often forget that they are creating something for people to actually use!

We discuss some of the lightweight usability testing methods, namely heuristic evaluation and user testing. These methods are simple, cost effective and should be a part of any design process for a system that involves humans. Spending a little time to ensure that the users - yes, the people who are actually going to use the system - are happy will save time and money and will result in your users wanting to use the system!

The basicsĀ 

Usability testing essentially involves evaluating the performance of a system in terms of its:

 

Heuristic Evaluation

In usability heuristic evaluation is a method of applying a set established heuristics (rules-of-thumb) to a system, to see how well it measures up to the standard. It aims to find those areas of the system that don't perform so well so that they can be fixed (hopefully before they are a problem). We have compiled such as set of these

Heuristic evaluation is best performed by usability experts, and domain experts but can be performed by others. Thus the downfall is that the results of the evaluation are highly dependant on the knowledge and ability of the experts, and that these experts may introduce extra (consultancy) costs. The advantage is that heuristic analysis can be performed quite methodically and quickly.

Heuristic evaluation is something that can be done during any stage of development. The advantage is that even before implementation, prototypes can be evaluated and modified before time and money is spent on its concrete implementation. It can also be used on existing systems in order to identify components that can be re-used and components that need to be re-designed in the new system.

User Testing

User testing involves testing the representative users of the system, preferably in their natural environment to get an idea of the actual use of the system. users should suffice in most cases, but of course testing as many different employees from different backgrounds, departments etc. will not be of any harm.

The actual test should be written up professionally and treated as any other project document. To conduct the test, following these three guidelines:

Where

Depending on the budget and resources allocated, usability testing may be carried out in a seperate testing laboratory, in the actual workplace in which the system will be used or a quiet office or conference room.

An important consideration to note; any testing requires that extraneous variables be minimised so that you have greater confidence that you have isolated the intended phenomena to be tested. However, testing the system in its usual environment might not be a bad thing - users won't in all cases be given a quiet or isolated room to learn your system but may have to use it in a busy, distracting environment. Therefore when testing in a quiet environment you cannot be sure that performance will translate accordingly to another, possibly distracting environment. Perhaps your system requires spatial or verbal tasks that seem simple but are actually easily affected by the presence of others. At the same time, your presence (the tester) might actually be enough to change the natural environment in which you thought you were testing leaving us with a lovely catch #22. Video and voice recording might be a possible alternative to your presence.

When

Usability testing should be carried out through each stage in the development life-cycle, as deferring it until the end is too late.

A Lightweight Usability Testing Regime

The following is an example guide of the stages that could be involved in usability testing: