alt Hacker NewsI was hired as employee 500 and now we are at 5000 (numbers rounded to protect the innocent)
It’s really really interesting what happens - obviously 5000 doesn’t compare to the behemoth that Microsoft is, but it’s still been interesting to see
It really is a confluence of factors that slows things down and I don’t think it’s as simple as saying business process
Some off the top of my head observations:
You have to try really hard to not settle on hires - the more you grow, the bigger the pile of work and you just want help - to the point where you might hire someone who’s good enough vs holding out for someone exceptional
It’s one thing for 100 people to be throwing money at problems in terms of tooling and outside contractors, once you’re big buying things is way more scrutinized - this is good because you don’t waste money on stupid tools, bad because it slows down buying of non stupid tools that you actually need
Things that didn’t matter before start to matter - security things, legal things, privacy things - these activities take time and also slow things down
Some work just doesn’t scale well - where in the past you could get away with throwing cheap manual labor at the problem, at some point you have to build automation to take care of - this speed things up in the long run, but usually you hit a wall first - as in things get so big they get slow - before you finally build the tools. Knowing which tools are worth building ahead of time is really hard
Related to the above - where in the past it made sense to buy a lot saas apps to run your business, these tend to be expensive so you start building your own
That’s just off the top of my head and nothing scientific. Also, I didn’t mention it specifically but of course you still have your run of the mill bureaucracy that slows things down
> Also, I didn’t mention it specifically but of course you still have your run of the mill bureaucracy that slows things down
Parallel friction: the more peers your work impacts, the more peers their are that can gatekeeper your work, for good and ill.
Grows roughly: log P, where P is impacted peers.
Serial friction: the more significant a change, the more management layers decision making has to pass up through, and every layer already has enough to do without considering significant changes. Can be very hard for a great idea to actually make it to someone who can say "yes", and for them to have the time.
Grows roughly: S^2, IMHO, where S is layers of "too busy" management to be traversed to a potential "yes".
That's the natural default. Some large organizations fight hard to keep the friction down, but it's going to be a fight.
After many years of nimble progress, a company I worked with went through a phase transition, and I spent as many years and more, slowly sinking into quicksand. Before giving up on my own project and moving on.