In Defence of Systems Thinking

I read “Magical Systems Thinking” by Ed Bradon this week. It is sharp, readable, and I agree with a lot of the caution. Complex systems tend to kick back, and it is wiser to grow from small working nuclei than to design grand contraptions on paper.

I still felt uneasy because the essay knocks down a version of “systems thinking” that many of us do not recognise. It treats the field as if it promises prediction and mastery. That strawman is easy to defeat.

The tradition some of us learned from sounds different. Donella Meadows ends her classic “Leverage Points” not with mastery, but with humility. She argues that the deepest leverage sits in mindsets and paradigms, and then goes one step further: real power looks like “strategically, profoundly, madly letting go.” That is epistemic humility in plain sight, not magical thinking.

A few other places where the piece felt selective to me:
> Treats Forrester’s World Dynamics like a failed forecast, not a scenario tool. That misses the point and the enduring insights about growth and feedbacks.
> Cherry-picks “failures” that are long-horizon and political, then contrasts them with “successes” that were crisis programs with narrow scope, top cover, and blank-cheque resources. Not quite apples to apples.
> Frames the alternative as “start simple” and “bypass bureaucracy,” but skips the relational work that actually makes simple systems take root: trust, shared expectations, and learning in practice.

So my takeaway is both/and. Yes, beware magical plans. Start from simple, working slices and evolve. And also remember what Meadows already taught: systems are counterintuitive, information is partial, people are not components, and humility is a method. The work is to stay in relationship with the system, learn fast, surface goals and rules, and only then abstract into structures that can hold.

Do you have other pieces that hold both the critique and the humility? I’d love to read them.

Originally written for LinkedIn on 16 September 2025. View original →