Methodologies & SDLC Deep Dive · 6 of 10

Lean — Eliminate Waste, Optimize the Whole

The Poppendiecks adapted Toyota's Lean Manufacturing to software in 2003. Where Scrum says "iterate," Lean asks "what's getting in the way of value reaching the user?" — and tells you to remove it. Lean Startup's "Build–Measure–Learn" loop is the same DNA pointed at product discovery.

WastePullBuild quality inDefer commitmentLean Startup
← Back to Methodologies & SDLC
The Seven Principles

The Poppendieck Set

  • Eliminate waste — anything not creating value for the customer.
  • Build quality in — automated tests, code review, refactoring; not "test it at the end."
  • Create knowledge — documentation, learning loops, dual-track discovery.
  • Defer commitment — make irreversible decisions as late as responsibly possible.
  • Deliver fast — short cycle times surface mistakes before they compound.
  • Respect people — empower the team that actually does the work.
  • Optimize the whole — don't local-optimize one team's metric while hurting the system.
The Seven Wastes (in software)

What to Cut

WasteWhat it looks like
Partially done workCode that's written but not merged / shipped / used.
Extra features"Just in case" code; the YAGNI smell.
RelearningSolving the same problem twice because nothing was written down.
HandoffsTickets bouncing between specialists / teams / approvals.
Task switchingMultiple in-flight items per person.
DelaysWaiting for review / build / deploy / answer.
DefectsBugs that escape and require rework.
Lean Startup

Same DNA, Different Domain

  • Minimum Viable Product (MVP) — smallest thing you can ship to learn.
  • Build–Measure–Learn loop — turn ideas into product, into data, back into ideas.
  • Validated learning — distinguish "users say nice things" from "users actually pay/return/share."
  • Pivot vs persevere — the explicit decision after each cycle.
Tradeoffs

What to Watch Out For

  • "Eliminate waste" can mean cutting muscle. Some "waste" — onboarding, training, slack — is investment.
  • Lean is a culture, not a checklist. Hard to retrofit onto a target-driven org.
  • Defer commitment doesn't mean "never decide" — late deciders still need to decide.
  • "MVP" gets stretched into "first release" — a real MVP can be a landing page or a Wizard-of-Oz mock.
Continue

Other Methodologies