Facebook, Google, Amazon et al succeeded _despite_ being bad/unknown quantities on paper. They specifically should _not_ be used as examples of "how to do it right."
Look at smaller sustainable startups that are flying under the radar to unearth proven, repeatable best practices. You can't establish patterns in methodology by citing exceptions.
For a $10mm company to become a $1bn company, know what you need? Luck.
EDIT: To those below, you're both validating my point. Luck isn't the sole driver of success, but it's the difference between a modestly successful company and an insanely successful company. The other skills the founders have to exploit opportunities are already there. The x-factor is luck. Nobody says "okay, all we need is 500 million users and we're successful!"