I have to admit that I’m not much of a focus group fan. They tend to produce such wonderful ideas as New Coke and the Bob user interface. Remember those gems?
It’s been a few years, but I’ve sat through a few focus group sessions during my career. I think the general problem is that focus groups tend to describe what they think they want, not what they really need.
A former employer of mine conducted a focus group in the early ‘90’s, with the goal of discovering more information about the typical computer book user. One of the things we were trying to determine is whether or not customers would pay more for a disk in the back of a book. The focus group results were unanimous: no way would these people pay more for a book with a 3-1/2” disk in the back. According to the feedback, it didn’t matter whether that disk included source code, utilities or the meaning of life itself – it just wasn’t worth a penny.
Interestingly, one of the imprints at that former employer was conducting the same test, only they were doing it with real products in real stores. They produced two versions of the same book: one came with no disk while the other was $5 more and included the disk. Guess what? The one with the disk typically outsold the other version by a factor of 2 or 3. It wasn’t just one title, by the way. They conducted the same test with no less than 10 titles over the course of a year. So much for the focus group results.
If focus groups aren’t that effective, what’s the best way to figure out what the market needs? Rather than trying to determine this in a simulated environment, why not just go directly to the customers at work or wherever they’re likely to use the product? If you’re working on a book about Excel, that means talking with people while they’re working with spreadsheets at the office. If the topic is Java, spend some time with a developer while they’re actually designing and writing the code. This isn’t as glamorous as a full-fledged focus group, but I’m willing to bet you’ll get better information in the long run.