When To Introduce Usability

The accepted wisdom that usability engineering should be introduced as early as possible in the development process has many arguments in its favour. Some of those arguements are outlined here.

On the basis that user-centred design, user testing and other forms of usability engineering are a 'good thing', it is worth asking how they are best integrated into the development process. Specifically, the issue of timing is critical to the successful implementation of a usability strategy. At what stage of development should these techniques be introduced?

Of course the answer to that question rather depends on the unique circumstances of each particular project, but the general principle seems to be "as early as possible" - with the important caveat that any user involvement is better than none, even at a late stage in the day. But the accepted wisdom that usability engineering should be introduced as early as possible in the development process has many arguments in its favour:

Involving users early - ideally before any development has taken place at all, establishes a valid base for all further development. If user input takes place early and often, the problem of investing significant time and money in solutions that fail to meet basic needs can be avoided. Early input means that if revisions are necessary, they are only small and will not involve the total redesign and development of an entire application or site.

User testing and input is often far more effective on early prototypes than more 'developed' interfaces. This is partly due to the greater willingness of test participants to criticise rough, easily changed drafts rather than 'finished' work. Early testing also enables prototypes to be changed quickly (or even on the fly) and then presented for further testing.

Usability professionals do not work in isolation, and early involvement tends to improve relations with designers, developers and project managers. Bringing testing in late - alongside QA testing for example - is inviting conflict with those whose work is criticised and management who have release deadlines to meet. Working together from the earliest stages will ensure that awareness is developed over time. In many cases user-centred design will actually speed up development if introduced sufficiently early.

But despite all these arguments, it must be remembered that something is always better than nothing. Although testing is always preferable before launch (in order to avoid early-adopters being disappointed and not returning), there is always a role for user testing on finished sites, or even those currently in operation. On the basis that many problems are often simple to resolve, it is always worthwhile employing user testing techniques to identify these issues. Sometimes the simplest changes can make a world of difference.