It is MUCH easier to create something brand new and acquire users than it is to get users to leave a similar product and use your instead.

I don't think that's necessarily true.
Bitchute is pulling in new users from YouTube by the thousands every month.
Parler (twitter equivalent), has pulled in over 500,000 !!! users in this last few days..
(Steem) Hive is flawed in much deeper ways than user interface issues, in my opinion.
A tinkering with the wing mirror, while the engine isn't running properly, kind of thing.

This exerpt is from @practicalthought, (in a comment on my post, @lucylin/would-you-care-to-dance-or-are-you-just-greedy)

....sums it up nicely.

...you mention onboarding one person per year. It could never happen. As you point out, 10000 would become 80000 within 4 years, which all things remaining constant percentage wise in activity, would dilute/reduce current payouts by almost 90% for most ( I say most as those on auto voters would still retain a higher share of the pool, but would still feel the impact a lot nonetheless).
This would also be exacerbated by the slowdown of inflation. Sadly, I do believe that it was probably known from creation that the math would only be sustainable if those in at the beginning held most of the initial stake and that instilling the idea that down voting was good for the community to stop theft and abuse would help extend the life of this illusion the math is sustainable.
A final observation on the intent is this. If the real objective was to bring in the masses, there would be some sort of reverse curve that rewarded lower staked accounts greater than higher staked, in order to promote ownership feelings and entice masses to come. As it stands now, most come, get little to no attention and leave. Or, they come with a large following, get the crap downvoted out of them to make sure they don't stay and reveal that bad reward math.