CF Dictionary · Theory of Constraints Terms
Variance (ToC)
Statistical fluctuations in real-world performance. ToC designs for it; CF treats it as a normal feature of systems.
In ToC, variance is the statistical fluctuation in real-world performance. Machines break down. People are late. Materials vary.
Variance ≠ error
ToC distinguishes variance from error:
- Error. A mistake; something done wrong.
- Variance. Natural fluctuation; nothing is "wrong", things just vary.
CF's interest: errors are correctable; variance is managable.
Why variance matters
- Real systems vary. Treating them as deterministic is wrong.
- Variance hits bottlenecks hardest. A small slowdown at the bottleneck costs more than a small slowdown elsewhere.
- Variance accumulates. Without protection, downstream effects grow.
ToC's variance management
- Buffers protect bottlenecks from variance.
- Margins are designed tolerance.
- Subordination. Non-bottleneck steps should not run at 100% utilisation — they need slack to keep the bottleneck fed.
CF's adoption
CF treats variance as a normal feature of life:
- Skills vary. Performance fluctuates with sleep, mood, etc.
- Outcomes vary. Even good plans sometimes fail.
- Design for variance. Buffers in habits, practice for resilience.
"Variance: Statistical fluctuations; errors that happen when things don't go perfectly according to plan." — LessWrong summary of CF