User>Driven

Jacob Nielsen published an excellent article way back in 2000 suggesting as few as 5 users were enough to adequately test a design. He shows data suggesting you gain 85% of the information it's possible to gain in user testing of a single design with 5 users and that more users provide less and less unique feedback. After that, he says, you might as well revise the design and start with 5 new users. Revise one more time and test with 5 more and you probably have as good a design as you are going to get.

I liked this article not just for it's pragmatic approach to research but also because I think it shows how easy and approachable doing product research really is. Companies often shy away from market and user research, I think, because it sounds intimidating. You need to sponsor customer advisory boards, organize focus groups, compose elaborate questionnaires, and so on, it seems. That sounds hard, takes a lot of time and money and requires expertise. Well, I'll agree with one of those.

In fact, you can learn a lot by just putting your designs in front of a handful of customers and asking them to try them out. You can also call up a half dozen customers and ask them about their biggest frustrations (within your are of expertise) and get a lot of information to feed into your feature prioritization.

I do think you need some expertise to do this right, however. As I have argued recently, someone in the organization must be dedicated to listening to the market - to customers, potential customers and users - to find out what the opportunities are for solving problems and whether proposed solutions are likely to solve them. These people can't be salespeople or sales support people. They will usually be product managers and design or usability professionals.

Once you've invested in hiring these people, though, you don't usually need an enormous research budget to support them. You can get the qualitative input you need with just a few customer interactions. (Though quantitative market sizing data may take a little more effort - see my recent entry on this.)

Is 5 subject enough in your experience? What's the fewest number of subjects you've used and felt like you met your goals? What is the largest number you ever used and why? Post your comments below.