Alan Cooper on Innovation

I saw Alan Cooper speak tonight at IxDA-SF.

Contrary to the popular business axiom of the first mover advantage, Cooper points out that nearly all successful products and sites are not the first to market-- not the innovators in their market. Usually the first product in a market is supplanted by a better designed, more carefully designed copycat. (This is actually Microsoft's standard business strategy- embrace and extend)

Cooper gave these examples:

• newton - palm
• some noname MP3 player - ipod

to which I'd add these examples

• netscape navigator - IE
• xerox star - mac
• friendster - myspace

Thus he concludes that Innovation does not lead to success. Rather, harnessing innovation matters. Notably, allowing sufficient time to get it right - get the design quality sufficiently high, is what ultimately leads to success.

Cooper also points out that software development is a craft, like carpentry, where quality requires time and experience (even apprenticeship), and cannot be schedule driven. This is further evidence that rushing to market generally will mean shipping a shoddy product which is doomed to failure for that reason, regardless of competitive dynamics.

Another Cooper observation is that programmers fit into two categories. Those that want to get something done and shipped (getting satisfaction about delivering profitable product), and those that want to craft the perfect technology solution for an abstract theoretical problem. These two types fit into Cooper's recommended two stages of software development-- design engineers (the latter type) and production engineers (the former type). He suggests that design engineers are suited to working with interaction designers to closely work to design both user experiences and technologies and frameworks to enable implementing them, while production engineers should be recruited to rapidly build and ship software to those designs.

In my observation, perhaps because of its older and more educated employee demographics, Sun is densely populated with engineers who cherish design engineering, and has fewer production engineers who just want to ship code. This has some interesting implications...