Shipping more code does not necessarily deliver value sooner. Reviews, test pipelines, deployment handoffs, and production feedback form a delivery system; speeding up coding alone may leave its delays and rework untouched. Lean is a way to examine that whole system, not a set of ceremonies or a demand that developers work faster. section 1.0 section 1.4
In software, value is a capability users can rely on over time—not a line count or a feature marked complete. A change must work in production and remain usable, reliable, and maintainable. The value stream is the path from an idea to that running capability, including the queues and feedback loops that are easy to overlook. section 1.1.1
Five principles turn that view into questions for day-to-day decisions:
- Define value: What user capability will this change provide, and can the team sustain it in production? section 1.1.1
- Map the value stream: Where does work wait, loop back, or lose context between intent and operation? section 1.1.1
- Create flow: Are review queues, slow tests, or handoffs delaying feedback and forcing work to restart? section 1.2.2
- Establish pull: Is the next step ready? Clarify constraints before implementation, and use clear readiness signals for review and deployment. section 1.2.2 section 1.2.5
- Pursue improvement: Surface defects earlier, then change the conditions that produced them instead of fixing only the latest incident. section 1.1.1 section 1.2.2
A three-line fix that takes ten days
A customer reports duplicate results. Investigation, fixing, testing, and documentation take roughly eight hours, but the fix reaches production ten days later. Triage, review, staging, and deployment each have a purpose; together, their queues, handoffs, and late feedback dominate the elapsed time. The Lean question is not which person should work harder. It is which delays the delivery system creates and whether they can be reduced without losing necessary safeguards. section 2.1
Do not label every activity that fails to change the product directly as waste. Tests, monitoring, and security scans may protect value even though customers do not experience them as features. Classifications depend on context and risk. To practice, audit a recent pull request: list the work, distinguish value-adding work from necessary support and avoidable waste, then ask which system constraint caused the largest delay. Start with one small change and check whether it improves the outcome. section 2.3 section 2.6 section 1.3


