There's something very attractive, particularly to developers, about the idea that product features should be driven not by what the business thinks it needs, but by what the developers want to implement. Scott Chacon of Github posted this presentation which suggests that Developer Driven Development can be successful under the following conditions:
- You must own your product
- You can have no deadlines
- You must have passionate developers
- Your team must have great communication, both online and offline
- You must have small teams
- The team should be users of the product
Scott suggests that this is a more successful way of engaging your development team, applying the open-source development approach to your company's product.
As I said, it's a very seductive idea to developers, and it certainly will work for some products, but I would strongly caution against applying this without a very good understanding of what makes it work and what makes it fail.
Product roadmaps are not inherently evil, and a great many companies have succeeded with them. Github was born in very special circumstances (zero credible competition, for a start). And, the three core developers on Github were also the founders of the company, and had an intimate understanding of its business objectives. In many other companies, this model would spell disaster.
If you read this far, you should follow my RSS feed here.