Feedback Loops

Why does growth that accelerated for a year suddenly flatten without anything breaking?

The idea

Systems change through reinforcing and balancing feedback.

In the real world

More users create referrals until support capacity constrains growth.

Going deeper

Reinforcing loops accelerate: more of something produces more of it. Balancing loops resist: growth in one place triggers a constraint that pushes back. Most systems contain both, and the observed behaviour is which dominates at a given moment.

A year of referral-driven growth that plateaus while spend is constant is usually a balancing loop engaging. If support times tripled, worse experiences are producing fewer recommendations, and the limit sits in delivery rather than acquisition. Spending more on the top of the funnel cannot move a ceiling created further down.

Where it stops applying

Identifying the dominant loop is harder than the framework implies, and multiple loops with different delays produce behaviour that looks like neither. The value is in asking what pushes back, not in a precise model.

Why it matters

Reinforcing loops explain the acceleration and balancing loops explain the ceiling, and both are usually present.

Try this today

Map one loop and its delay.

Test yourself

Referrals drive growth for a year, then plateau while spend stays constant. Support response times have tripled. What is the relationship?

Show the answer

The reinforcing loop of users creating referrals is now being held down by a balancing one: slower support produces worse experiences and fewer recommendations. The limit is not in acquisition, so spending more there will not move it.

Learn this in the feed Answering from memory, then again days later, is what makes it stick.

More in Product & Systems Thinking

Small Experiments Local Optimisation Hurts Design for Failure Bottlenecks Govern Flow Queues Reveal Capacity Observability Enables Learning

All Product & Systems Thinking lessons