Having worked at Amazon for almost 4 years before moving to a start-up, I always find it very interesting to talk to other large-company ex-employees who bemoan the stiffling and slow-nature of process. For example, I just talked with an ex-materials engineer who was so frustrated with the speed in which it took him to implement or accomplish anything at his company that he left to go back to school. I also remember talking to an ex-employee of a government contract industry who hated his job so much he quit before he lined up his next gig. Things moved too slow and ultimately he (they) never felt like he was doing anything.
Why do companies have process, to start with? Simply put, process defined as “a series of actions or steps taken in order to achieve a particular end.” In short, this is an agreed upon set of steps for people to take in coordinating and trying to accomplish something.
In contrast and in no surprise to anyone, most startups have very little process. The lack of process in a very small company can be very, very empowering. This means a team of two engineers could theoretically design and create an entire web-app without going through approval with marketing, sales, management. This proposal and review process is in itself is a full time job that requires considerable writing, communication, prep, and review.
My current startup has grown from ~8 people only a year ago to now approximately ~50. As a little context, as a software tech company, our company is compromised of approximately ~70% engineers/research-scientists and ~30% other including sales/ops/marketing. The bulk of of 50 people (~35) are developers split apart into 4 different teams with slightly different goals. We lack a lot of process but are actually trying our best to implement some changes.
Why? Without process, it is very, very difficult for such a large group of people to coordinate and do their job. In my eyes, this breaking point happens when there are 2 slightly dependent teams (ie. product readiness before sales can pitch) OR the team grows beyond ~4 people. For example, I product manage a dev team of ~10 engineers and since they rely on me to give a vision, requirement, and plan, I need to give them information in a format they can consume. Engineers need to communicate to me how much work a certain project is and progress on other development, respectively. If we don’t work together in a set process with a timeline/format of communication that works for both of us, they’ll build something I don’t want and I will deliver ideas that are never taken or implemented. Without a good process, things can explode. As one example, sales sold a product to a large client in which the engineering team was not originally planning to finish. As a result, the development work had to be expedited into the roadmap and schedule, which led to subsequent delays and drops in other incredibly important features. Without a good production deployment process, the Q/A team can miss testing massive features and the company can push out a completely defective product (which just sucks and can be a trust-buster with existing customers).
When I was a Product Manager at Amazon, one thing that I took for granted was total buy-in to the process. Process did slow down some teams, but I felt blessed to work with some vision-focused leaders who would cut through trivial process to make progress against major goals. Non-surprisingly, it can be very difficult for an early-stage start-up engineer, who is accustomed to Rambo-ing to the beat of his own drum, to suddenly need to review plans, communicate timelines, and focus on cross-team coordination. However, the goal is that a well coordinated large team can do far more than one individual can.
Process prevents massive screw-ups. As the company gets bigger and the stakes get higher (ie more to lose with existing customers, more eyeballs, and more dollars at stake), it is more important to avoid massive screw-ups than it is to experiment
Big companies generally already have strong product market fit, as well, and a strong process goes hand in hand with established revenue streams. If there is no strong product market fit and a company implements process too quickly and too aggressively, I strongly believe that the lag in experimenting (and many times what may actually seem reckless) will render that company so slow that it risks never actually achieving product market fit fast enough.
While process can over-correct and stifle employees into a Taylorism assembly line robot, some process can help a company beyond a certain stage (ie. read as big enough, with enough different initiatives and functions!) coordinate and become a much more effective company .
 Even with unorthodox process in companies like Zappos, there is still a process in which almost everyone agrees to adhere to.