Donella Meadows’ list of places to intervene in a system has one line that has been sitting with me for weeks: people know intuitively where leverage points are, and then push them in the wrong direction. Her example was Jay Forrester’s finding that low-income housing needed less of it built downtown, at a time when the entire policy apparatus was pushing to build more.

I see a smaller version of this constantly at work. Teams add another approval gate to fix a quality problem, which is a parameter change, low on Meadows’ list, when the actual issue is a missing feedback loop, information not making it back to the people who could act on it. Adding gates without adding feedback just slows the system down without making it smarter. The fix that would actually work is usually higher up the list and much less comfortable to propose, because it means changing who has access to information, not adding another checkbox.