Another hint: coercion is not necessarily direct violence.
Capitalism is almost completely coercive and controlling, but direct violence is mostly reserved for minorities and for war adventures in distant countries.
At home it's coercive through financial violence (especially income restriction and forced debt), extreme economic apartheid, and overt and covert behaviour modification narratives in the various media.
"The crowd" is very much not in charge of any of this, but is persuaded that it is.
Arguably all of this is better than a more overtly dictatorial system where any form of dissent attracts direct violence. But it's not necessarily less coercive - just less obvious. And more efficient.
You can scale that back and have collective systems which rely on a tradition of consensus. It's not even all that unusual on a smaller scale. Teams/groups make decisions, some people disagree, but when consensus is reached everyone works towards the same goals even if they have misgivings.
This works if you allow feedback, goal monitoring, and resets so the system is responsive and self-correcting. But it also requires all participants to be rational, reasonably intelligent, non-sociopathic, and capable of adult decision making.
Which is why it doesn't work at scale. Only about half the population - at best - has those abilities, and as soon as your group includes at least one sociopath they will destroy that dynamic and replace it with a toxic one which relies more on coercion than consensus.