As we saw, Conway’s law states that you cannot design architectures that differ much from the organization’s communication’s structure.

Clayton Christensen explains that big companies have a hard time innovating (read The innovator’s dilemma). And that most innovation happens outside the walls of a big company.

Actually, both ideas are related. Since, to innovate you need to be able to be disruptive. Being disruptive means that you -among other things, may- need to change your architecture(s) radically. But if you cannot make other architectures than the ones that closely reflect your company’s communication structures, you won’t be able to create innovative products.

So your only way out is making incremental changes to your architectures that only deviate a little from the company structures (as to not shock anyone). But this iterative approach takes time. A lot of time. Certainly compared to a start-up that can start with a radically new architecture from day one.

And this is a good thing, because while the big companies are iterating towards a better architecture (or are not doing any changes at all) the start-ups can take a lead and grow without fear of being immediately taken over by a bigger company.

By the way, mr. Christensen gave an interesting talk at Business of Software 2011.