One solution to this dilemma that we have seen successful in this situation is the "Skunkworks". Basically, you collect a group of excellent solution developers and marketeers led by a passionate champion and unfetter them as much as possible from burdensome company processes. Creating a Skunkworks successfully is very challenging. Entrenched managers will resist giving up budget, resources and authority. Bureaucratic gatekeepers will resent results that bypass their carefully constructed processes. And, if not done carefully, you can create an unhealthy status competition and resentment between workers in the Skunkworks and those in the "normal" groups of the company.
The main point is to let one part of the company work on protecting its brand and dominant market share while insulating another part of the company to go after new markets with high speed and unfettered innovation. The Skunkworks is one way to do this, usually as a temporary organizational measure when a clear opportunity presents itself. An alternative (perhaps growing out of a successful Skunkworks) is to spin-out a subsidiary company. Trying to accomplish both mature product market protection and entry into fast-changing new markets under one, matrixed organization with internal competition for resources, inflexible processes, and divided management mindshare is a recipe for failure.
