I wholeheartedly disagree that the cost is "nothing, nada, zilch". The network effects of having everybody using a couple of sites is what makes the cost of switching high, not the technical triviality of creating a new website.
The cost of entering the mass-scale social media market is basically zero, since you can do that in a day with minimal cost other than a domain name and a server (or VPS or AWS/Azure account or similar), and you can scale your costs as you grow, but your costs don't grow as fast as your customer base.
A natural monopoly's costs start high, and just scales from there. It costs tens of millions for a utility just to enter a small market, and with utilities, growing may actually increase their costs (see, for example, power and water utilities).
>The barrier to creating a new social network once network effects have locked in a few major players is enormous.
So...that is what I said, that the barrier to entering the market is enormous.
I'm definitely dating myself here, but Livejournal, Blogger, Myspace, Frienster, DailyBooth, Friendfeed, Yikyak, Snap, Vine, were all once the dominant social media website before losing to new entrants.
Facebook was started in a college dorm room. Twitter launched as an employee's side project. Instagram began with two roommates sharing an apartment. WhatsApp began as one guy.
There’s no physical issue preventing competition like with power companies, but to a social media startup, getting enough users is a similarly hard challenge as constructing a new power grid.
And you’re not actually a social network until you reach some critical mass of users. The cost of entering the market is low as you point out, but that doesn’t imply that you can realistically compete. So I think it’s fair to call social networks a natural monopoly. Even if they don’t strictly fit the definition they’re still monopolies in practice.