Theory of Constraints — the one bottleneck that limits everything else
18 Apr 2026
Theory of Constraints (TOC) was developed by Eli Goldratt in his 1984 novel The Goal. The principle: every system has exactly one binding constraint at any moment. Optimising anything else is waste — the system can only move at the speed of the constraint.
This is counterintuitive in a culture that values “improving everywhere.” Goldratt’s insight: improvements not at the constraint don’t actually improve the system. They just produce more inventory upstream of the bottleneck while the bottleneck still drips at the same rate.
The five focusing steps
- Identify the constraint. What is actually limiting the goal? Not what’s annoying. Not what’s most obvious. What’s the rate-limiter?
- Exploit the constraint. Squeeze maximum throughput from it without buying more capacity. Stop interrupting it, stop letting it idle, prioritise it.
- Subordinate everything else to the constraint. Other systems serve the constraint, not the other way around.
- Elevate the constraint. If exploitation isn’t enough, now add capacity (hire, buy, train, build).
- Repeat. A new constraint will appear elsewhere. Find it.
The error most people make: jumping straight to step 4 (add capacity / try harder) without doing steps 1–3.
Examples in life, not factories
Most people optimise what’s easy or visible. The actual constraint is usually painful and avoided. That’s why it’s the constraint.
| System | Common false constraint | Actual constraint |
|---|---|---|
| Health | ”I need a better workout program” | Sleep |
| Wealth | ”I need more income” | Spending or interest-bearing debt |
| Career | ”I need new skills” | Visibility; the conversation you avoid |
| Relationships | ”We need to communicate better” | The thing that’s never been said |
| Creative project | ”I need better tools” | Shipping anything |
| Estimating accuracy | ”I need a better template” | The slowest contractor’s quote |
The constraint is not where you wish it was. It is where the system actually slows down.
Subordination is the hardest step
Once you identify the constraint, everything else has to serve it. That means:
- Stop optimising secondary systems. They’re already faster than the constraint. More speed there is wasted.
- Stop adding new initiatives until the constraint moves. New projects compete with the constraint for resources.
- Tolerate slack in non-constraint resources. Idle capacity at non-constraints is normal — it means you have extra room. Idle capacity at the constraint is catastrophic.
How to find your constraint
If you’re not sure what your constraint is, ask:
- Where does work pile up? Inventory accumulates upstream of bottlenecks.
- What do I avoid talking about, addressing, or measuring? We avoid what hurts. The hurt is usually the constraint.
- If I could fix one thing instantly, which would unlock the most? Forces ranking against the goal.
- What’s the same problem month after month? Persistent issues are usually constraint-level.