I’ve been a massive fan of Steve Krug for years. I discovered him during a brief detour in my career when I was working as a web designer. I bought a book called Don’t Make Me Think which was an introduction to both web usability and to the process of usability testing. I realised that most website were badly designed because most web designers never showed their work to anyone before it was released. Krug argued for an iterative and consumer informed approach to design. In essence you have an idea – you show it to some people, you build a prototype – you show it to some people, you get your first page up – you show it to some people and so on.
One of the things that I really remember from that book was the way that Krug talks about how the idea of doing things properly stops them from happening at all. So because you can’t find the full demographic spread of users you don’t show your design to anyone. Or because your site is meant to be viewed in schools you hold off on showing it because it is difficult to get into schools. Krug makes a simple point that chimes with me. The more you show your work to other people the better. The more the people you show it to are like your intended audience the better, but don’t let this stop you showing it to someone. This principle works for website design, but it also works for pretty much everything else (academic articles, your CV etc.) Of course there are important qualifiers to this. Some sense of representativeness should be an aim of evaluation projects, but Krug’s point is essentially that the perfect is the enemy of the good. One brilliantly representative evaluation of a product once the product is designed is probably a lot less effective than 10 imperfect evaluations as you develop the product.
Anyway, this post was prompted by chancing across this video of Krug explaining how to run a simple usability test.