Being involved with a school provides a good example of scaling problems. A lot of things that seem intuitive or simple at an individual level are difficult or complex at larger scales.
One key number is 30, the approximate number of children in a class (Helen's has ranged from 28 to 31). The other is 450, which is roughly the number of children in her school, a two-form entry primary school with an attached Early Years unit.
Back in Year 1 I can remember discussing with parents what kind of feedback we expected on the children's reading diaries. At first I thought it would be nice to get (say) weekly comments by the teacher on these. But with thirty children, providing non-trivial comments here could easily take the teacher two hours a week, which when I thought about it probably wasn't the best use of their time.
Want to give everyone in a year something? That's 60 children, so gets expensive pretty quickly. I managed it with books in Reception, but only because I found two cheap copies of the thirty book Walker Stories set. (I've found other sets of books cheap enough to do this, but none that I liked enough to want to give to everyone.)
At a school level, scaling issues come up with provision of food, where even marginal increases in the cost of a lunch rapidly impact on budgets. And during the coronavirus closure, our school has had teachers ringing all the children weekly. That doesn't seem like much, but even at five minutes per call, with 400+ children that's over 30 hours a week, the equivalent of an entire teacher working full-time.
The converse of this, however, is that if you can do something to help a school as a whole — provide reusable resources that everyone can draw on, get training for the teachers, improve morale and esprit de corp, improve the buildings or playgrounds, etc. — then that's a pretty direct way to improve life for a large number of people.
And the huge economies of scale schools offer are obviously a gain in many areas. As a lot of us have had brought home to us in the last few months, it's a lot more efficient having thirty children taught and minded at school than individually or in family groups at home!
For an extreme example of scaling, Facebook makes approximately $3 per year profit.
That's $3 per Facebook user per year profit, of course, which turns into over $7,000 million per year for the company as a whole.
This has the same consequences as you describe for the school but scaled - if the cost of doing something is a one-off for the entirety of Facebook, there's huge money available to fix it. If the cost of doing something scales per-user, then it's basically unaffordable.
Or, to put it another way, if the only way to fix a problem is to invest $5 per user on Facebook, it's uneconomic. If you can fix it by investing $1,000 million per year in a system that fixes it no matter how many users there are, it's fixable.