Yes, 1 million users is an Everest - and collapsing at the top you find it is merely the foothills, above you lies Mons Olympus, and there's not even oxygen up there.
Yes, that is exactly the way I meant it. The post I referenced at the top on orders of magnitude (http://ye.gg/magnitude) phrases the same concept another way. You generally have to go through a lot of orders of magnitude to get to one million, and then everything that worked to get you there probably won't work to get you to ten million in any reasonable amount of time. Unless of course you have that amazing organic growth.
It feels like its not about scaling - I could design a system that could handle ten or hundred fold traffic increase and just need more servers and config.
But at some point I need to change - like I could have a career of the same years experience repeated twenty times. Or I could grow in each year and come out of a decade a different person
It seems that duckduckgo has same code as google just needs more servers - but as each year of experience or each order of magnitude passes it would not be enough to buy more servers - you would need to own fibre networks.
(Don't get me Wrong on "same code as google" - I am always amazed by how many developers it seems to take to do things - google for a company that "just" runs web crawlers and map-reduce seems to have about 10,000 too many staff.
Which I guess puts me in the camp of someone who would have ground google to a halt at 100 staff saying - we don't need to hire more people, just write code to do it instead.
So goin a little off the point there - but there is a metric of growth that is not "the same users repeated a million times" but is a new and different company - and you hit that at ten or a hundred. And if you are a twenty years repeated kind of person you won't grow past that level.