Quartz 4

Lean Flow Theory Applied to Software

2 min read

Core Principle

Lean flow theory, originating in the Toyota Production System and formalized by Reinertsen’s application of queuing theory to product development, is the theoretical foundation for continuous delivery, trunk-based development, and modern DevOps practices. The core insight: batch size is the single most important determinant of cycle time, and fast feedback is the mechanism by which quality improves.

Why This Matters

The lineage is: Toyota Production System (1950s) → Lean Manufacturing (Womack & Jones, 1996) → Lean Product Development Flow (Reinertsen, 2009) → Continuous Delivery (Humble & Farley, 2010) → Accelerate/DORA (Forsgren, Humble & Kim, 2018). Understanding this lineage gives you the theoretical vocabulary to explain why practices like trunk-based development work, not just that they do.

Evidence/Examples

Reinertsen’s three key principles from queuing theory:

  1. Smaller batches flow faster through a system, with less variability, generating feedback sooner
  2. Large batches create queues, and queues are invisible inventory with carrying costs. In software: open branches, pending merges, staging environments are all WIP inventory
  3. Fast feedback reduces uncertainty. The sooner you learn a change is broken, the cheaper it is to fix. This is an information theory argument, not just an efficiency argument

The Poppendiecks’ vocabulary for applying lean to software:

  • Waste: anything that doesn’t deliver value
  • Flow: smooth movement of work through a system
  • Pull: downstream demand drives upstream work
  • Amplify learning: short feedback cycles

Implications

  • Practices that increase batch size (long-lived feature branches, release trains, staging gates) degrade both speed and quality
  • Practices that decrease batch size (trunk-based development, feature flags, continuous deployment) improve both
  • The argument is mathematical (queuing theory), not just empirical, which makes it more portable across contexts

Questions

  • How does Reinertsen’s model account for cognitive switching costs? Smaller batches mean more context switches.
  • Is there a minimum viable batch size below which the overhead of deployment exceeds the feedback benefit?

Sources

  • Reinertsen, D.G. The Principles of Product Development Flow (2009). Celeritas Publishing.
  • Poppendieck, M. & Poppendieck, T. Lean Software Development: An Agile Toolkit (2003). Addison-Wesley.
  • Humble, J. & Farley, D. Continuous Delivery (2010). Addison-Wesley.
  • Womack, J. & Jones, D. Lean Thinking (1996). Simon & Schuster.

Graph View

Backlinks