A feedback loop is when the result of something comes back and changes the thing that caused it in the first place. That's the whole idea. If the change makes more of the same thing happen, it's called positive feedback, or reinforcing feedback. If it pushes things back toward where they started, it's called negative feedback, or balancing feedback. Neither word is a judgement. Positive doesn't mean good, and negative doesn't mean bad.
We used a version of this idea in an earlier piece, without naming it outright. In Why Linear Planning Fails in Complex Systems, we wrote about a city that laid pipes to reduce how far women and girls walked for water. The scheme ran dry often, so families kept using private tankers. The tanker operators, now with steady business, lobbied against the pipes getting fixed. The fix had fed the very problem it was meant to solve. That's a positive feedback loop, quietly working against the intervention instead of for it. It's exactly why feedback is a term worth actually understanding, not just nodding along to in a meeting.
Where You'll See It
The same loop shows up in systems that have nothing to do with each other:
A stampede in a crowd. One person panics, the people near them panic more, and it spreads outward faster than it started. Positive.
Sweating and shivering. Your body pushes back against a temperature change until it settles at 37°C again. Negative.
Melting permafrost. As the climate warms, permafrost melts and releases greenhouse gases, which warms the climate further. Positive.
A savings group repaying loans on time. Early repayment builds trust, trust unlocks larger loans, and larger loans, used well, build even more trust. Positive, and one worth designing for on purpose.
Look at those four examples again. A crowd panicking, a body cooling itself down, a warming climate, and a savings group building trust are not the same kind of system at all. One is about people, one is about biology, one is about the planet, and one is about money. But the pattern running through all four is identical. A result feeds back and changes its own cause. Once you can see that pattern, you start noticing it in places you never thought to look.
The One Thing People Get Wrong About It
In everyday use, "feedback" usually means a comment. What your manager says about a report, what a community says about a workshop. That's a useful thing too, but it's not what the term means in systems practice.
A feedback loop isn't someone's opinion arriving from outside the system. It's the system talking to itself: a result that becomes, in turn, a cause. This distinction matters in practice. When we're mapping a problem and ask where the feedback is, we're not asking who has an opinion. We're asking what part of the system is quietly reinforcing or dampening itself, often without anyone noticing until the pattern is well underway. That's exactly what happened with the tanker operators. They had already taken hold before anyone named the loop they were part of.
Key Takeaways
Feedback loops sit at two extremes. One can lead to runaway, accelerating effects. The other creates inertia, dampening change and holding a system in place. Most real systems are a tangle of both, pulling in different directions at once.
Positive feedback is reinforcing. It accelerates whatever direction the system is already moving in, for better or worse.
Negative feedback is stabilising. It suppresses change and pulls a system back toward balance. That is exactly what makes a system resilient, not stagnant.
The next time a solution seems to be working too well, or a problem seems unusually stubborn, it's worth asking the same question: what's feeding back into what here?
Where have you experienced feedback loops in your work? Tell us in the comments!
Related reading: Emergence: What It Actually Means · Why Linear Planning Fails in Complex Systems