Lean Startup

Lean Startup is a methodology for developing products under conditions of extreme uncertainty. Created by Eric Ries (building on Steve Blank's Customer Development), it replaces traditional planning with a…

Guide · Updated · Source

What it is

Lean Startup is a methodology for developing products under conditions of extreme uncertainty. Created by Eric Ries (building on Steve Blank's Customer Development), it replaces traditional planning with a Build-Measure-Learn feedback loop that maximizes validated learning while minimizing wasted effort.

The core insight: the biggest risk in product development is not building it wrong (an SDLC problem) — it's building something nobody wants (a PDLC problem). Lean Startup addresses this by treating every product idea as a hypothesis to be tested with real evidence, not a plan to be executed.

Authoritative sources (external)

Resource Executive summary (why it's linked here)
The Lean Startup — Eric Ries Canonical text defining Build-Measure-Learn, MVP, validated learning, pivot/persevere — the philosophical anchor for hypothesis-driven product development.
Steve Blank — Customer Development Precursor framework: Customer Discovery → Customer Validation → Customer Creation → Company Building — the business-model-validation layer underneath Lean Startup.
Lean UX — Jeff Gothelf UX integration of Lean Startup principles into Agile teams — hypotheses, experiments, outcomes over outputs. Bridges Lean Startup with design practice and SDLC iteration.
Running Lean — Ash Maurya Practitioner playbook for applying Lean Startup systematically — Lean Canvas, experiment design, metrics that matter.

Core structure

The Build-Measure-Learn loop

Linear flow diagram template

Direction of execution: Build → Measure → Learn. Direction of planning: Learn → Measure → Build — decide what you need to learn first, then what to measure, then what to build to get that measurement.

Key concepts

Concept Definition PDLC connection
Hypothesis A falsifiable statement: "We believe [action] will [outcome] for [audience] because [reason]." P1–P2: every experiment starts with a hypothesis
Minimum Viable Product (MVP) The smallest thing you can build/do to test a specific hypothesis. Not a "version 1" — a learning vehicle. P2: validation experiments
Validated learning Evidence that confirms or refutes a hypothesis — not opinions, not vanity metrics P2 exit criteria
Pivot A structured course correction: change one element of the strategy while preserving what you've learned Gate G2 "pivot" decision
Persevere Evidence supports the hypothesis — continue on current path Gate G2 "go" decision
Innovation accounting Measuring progress toward validated learning, not just activity P3 success metrics definition

Types of MVPs / experiments

Not all MVPs require code. Choose based on what you need to learn:

Experiment type What it tests Cost Speed PDLC phase
Problem interview Does the problem exist? How painful is it? Very low Hours P1
Solution interview Does the proposed solution resonate? Very low Hours P1–P2
Landing page / fake door Would users sign up / click to use this? Low Days P2
Concierge MVP Can we deliver value manually before automating? Low Days P2
Wizard of Oz Does the experience work if we fake the backend? Medium Weeks P2
Paper / Figma prototype Can users navigate and complete core tasks? Low Days P2
Coded MVP Does the full solution deliver value in production? High Weeks P2–SDLC

Mapping to PDLC phases

PDLC phase Lean Startup activity
P1 Discover Problem Problem interviews and Customer Discovery (Blank) — validate that the problem exists and matters
P2 Validate Solution Build-Measure-Learn loops: MVP experiments, usability tests, concept validation — validate that the solution addresses the problem
P3 Plan & Commit Innovation accounting: define success metrics, establish baseline, set targets that indicate product-market fit
SDLC A–F Build the validated solution. Lean Startup's "Build" phase for production (vs experiments)
P4 Launch Customer Creation (Blank) — test go-to-market channels, pricing, positioning
P5 Grow Ongoing Build-Measure-Learn: A/B tests, feature experiments, retention optimization. Pivot/persevere at product level.
P6 Mature / Sunset Pivot or end: evidence-driven decision to reposition or retire the product

Pivot types

When evidence says "don't persevere," these structured pivots preserve learning:

Pivot type What changes Example
Customer segment Target audience B2C → B2B for same product
Customer need Problem being solved Analytics → reporting (adjacent need discovered in interviews)
Platform Delivery mechanism Mobile app → browser extension
Business model Revenue approach Subscription → freemium + marketplace
Channel Distribution Direct sales → self-serve
Technology Implementation approach Custom engine → open-source integration
Zoom-in A single feature becomes the product Dashboard widget → standalone dashboard product
Zoom-out The product becomes a feature of something larger Standalone tool → integrated platform module

Anti-patterns

Anti-pattern Fix
MVP = crappy v1 MVP is a learning tool, not a bad product. It's the minimum thing that tests a specific hypothesis. Some MVPs have no code at all.
Vanity metrics Measuring page views, downloads, or sign-ups without connection to value delivery. Use actionable metrics: activation, retention, revenue per user.
Pivot avoidance Ignoring evidence because the team is emotionally invested. Set pivot criteria before running experiments. If criteria are not met, pivot.
Hypothesis-free experiments Running A/B tests without stating what you expect and why. Every experiment needs: hypothesis, method, success criteria, sample size.
Premature scaling Growing before validating product-market fit. "We need more users!" is not validated learning — it's hope.

Further reading

  • Design Thinking — Complementary: adds empathy-first problem framing before hypothesis generation
  • Opportunity Solution Trees — Visual structure for organizing hypotheses and experiments
  • Stage-Gate — Complementary: provides organizational governance for Lean Startup experiments
  • PDLC-SDLC Bridge — How validated learning crosses into delivery