I am in the midst of training for the GSA, Google Search Appliance. While training is usually fast and furious, it is nice meeting other people from different companies and seeing how everyone is viewing the same product. The different perspectives are nice to have, sometimes great for new ideas.
I was conversing with a fellow trainee and we were talking about why we were at the training. He comes from a company of about 50 people and me from 90,000+. Both of us got onto the topic of pros and cons about company size. Small companies are quick and agile with lots of collaborating, and there is usually little overhead. Larger companies have lots of resources and influence but usually have lots of bureaucracy, aka. overhead.
Many try to argue one is better than the other, but it is really what is suited to your tastes. I work for a company that has a lot of overhead, but it makes sense. Everyone has worked on a project where management does not step up and people start running in different directions. Doing things slowly right, is faster than doing things wrong really quickly and having to redo the mess up.
Personally, I am not a fan of people that are pure management. I know it is a necessary evil and I am not going to get rid of it. My major qualm is those people creating large amounts of unnecessary bureaucracy. Some people are so high up they cannot see the ground. The only information they receive is from other management.
Governance is extremely important and not fun to implement. I understand this, but if bureaucracy needs to be created it should be the job of those in charge to make sure that it runs as efficiently as possible. Governance is there to protect the business so things do not go wrong. If it gets to the point that it starts harming the business then I think there is a problem. People of my level have no way of influencing high level business policies, those in charge should be constantly trying to improve company inner workings the same way developers improve and refract code. Business process and computer code are in the end both algorithms.