It's worth remembering how Facebook got to the userbase that it currently has -- they didn't start out targeting "everyone on the planet", which is an almost impossibly hard challenge because of the network effects you note.
Instead, they started off providing closed social networks for existing groups of users -- colleges, starting with a small number of elite colleges and slowly broadening out from there -- until eventually it became a public social network and the de facto standard at least within the US.
The networks that have attempted from the start to get everyone to sign up have mostly failed. E.g. Google Plus, which had some feature advantages over FB at the time, started out much too ambitiously in my opinion and suffered from a "Potemkin village" issue where (unless you had a circle of friends who all joined it) it felt empty, and I suspect many users never used it more than a few times.
The only recent network I've seen to go the Facebook route is Nextdoor, which at least in my anecdotal experience seems to be becoming a thing. By providing a semi-closed social network based on physical proximity (neighborhoods), it provides immediate value to a new user. I'm not sure of the details of their rollout strategy, but at least at one point they insisted on mailing you a physical postcard in order to verify your address (much like FB's original validation on specific .edu email domains).
Anyway, if anyone out there is thinking of challenging Facebook on the social-network front, I would put some significant thought into the rollout strategy and aim not to compete with Facebook circa 2017, but instead to compete with Facebook circa 2004. As other networks have opened themselves up, there's a constant vacuum at the lower, more-specific, more-exclusive, closed end of the spectrum which provides a lower barrier to entry.