Redr · Study Guide
The Lean Startup
How Today's Entrepreneurs Use Continuous Innovation to Create Radically Successful Businesses
Eric Ries
Unofficial AI-assisted study guide. Not affiliated with or endorsed by the author or publisher. For educational use — supplements, not replaces, the original work.
Contents
- 00Introduction
- 01Start
- 02Define
- 03Learn
- 04Experiment
- 05Leap
- 06Test
- 07Measure
- 08Pivot (or Persevere)
- 09Batch
- 10Grow
- 11Adapt
- 12Innovate
- 13Epilogue: Waste Not
- 14Join the Movement
- Part 01 · Vision — A New Discipline for Entrepreneurship00Introduction01Start02Define03Learn04Experiment
- Part 02 · Steer — Build, Measure, Learn05Leap06Test07Measure08Pivot (or Persevere)09Batch
- Part 03 · Accelerate — Scaling Lean Without Losing Speed10Grow11Adapt12Innovate13Epilogue: Waste Not14Join the Movement
Part 01
Vision — A New Discipline for Entrepreneurship
Ch. 0–4
Introduction
Ries attacks the romantic myth that startup success comes from grit and a great idea, arguing instead that most startups fail because they lack a disciplined process for navigating uncertainty. The Lean Startup is presented as a scientific approach to entrepreneurship, drawing on Toyota's lean manufacturing and the lessons Ries learned the hard way at IMVU.
The Startup Myth
The popular narrative says success comes from determination, a brilliant idea, and being in the right place at the right time. Ries argues this mythology obscures the systematic causes of failure and gives entrepreneurs no actionable guide — leaving them to attribute outcomes to luck or charisma instead of process.
Entrepreneurship as Management
Startups are not just smaller versions of big companies; they operate under conditions of extreme uncertainty that break traditional general management. Yet they still need management — just a new kind, an "entrepreneurial management" suited to discovering an unknown business model rather than executing a known one.
The Five Principles
The book's structural backbone: (1) entrepreneurs are everywhere, (2) entrepreneurship is management, (3) validated learning, (4) build-measure-learn, (5) innovation accounting. Together they reframe startup work from heroics into a learnable discipline.
Roots in Lean Manufacturing
The methodology adapts Toyota Production System ideas — small batches, just-in-time production, eliminating waste, respecting worker knowledge — and applies them to knowledge work. Where Toyota fought physical waste, Lean Startup fights the waste of building things customers don't want.
The Biggest Source of Waste
The single biggest source of waste in product development is building something nobody wants. Lean Startup is designed to detect this fast and stop it before more effort is poured in, treating learning what to build as the central problem rather than execution speed.
- Lean Startup
- A methodology for developing businesses and products under extreme uncertainty, emphasizing rapid iteration and customer-validated learning.
- Validated Learning
- Progress measured by empirical evidence about what customers actually want, not by features shipped.
- Build-Measure-Learn
- The core feedback loop that turns ideas into products, measures customer response, and decides whether to pivot or persevere.
- Minimum Viable Product (MVP)
- The simplest version of a product that allows a team to begin learning from real customers as quickly as possible.
- Innovation Accounting
- A measurement discipline designed for startups, where standard revenue and ROI metrics start at zero and tell you nothing.
- Pivot
- A structured course correction that tests a new fundamental hypothesis while preserving the vision.
- Extreme Uncertainty
- The condition in which the customer, market, and right product are all unknown — the defining context of a startup.
Multiple choice
According to Ries, what is the single biggest source of waste in product development?
True / False
Because startups operate under extreme uncertainty, Ries argues they should be run without any management discipline at all.
Multiple choice
Which set correctly lists Ries's five principles of Lean Startup as introduced in the book?
Spot the issue
A founder explains startup outcomes by saying "you just need grit and a great idea — the rest is luck and timing." According to Ries, what's wrong with this view?
Start
Ries argues that startups urgently need a new management discipline — entrepreneurial management — because both "just do it" hacker culture and traditional MBA-style management fail under extreme uncertainty. He lays out the five principles of the Lean Startup and positions the methodology as a rigorous, learnable practice rather than a matter of luck.
Two Failure Modes
Startups fail in two opposite directions: (1) the good-plan trap — exhaustive market research and detailed forecasts designed for stable businesses, which collapse on contact with reality; and (2) the just do it school — throwing planning out entirely and grinding code in the hope something sticks. Lean Startup is the disciplined third path.
Vision–Strategy–Product Pyramid
Three layers change at different speeds. Vision is the unchanging end state — the world the founder is trying to create. Strategy is the plan to get there (business model, customers, partners) and is what gets pivoted. Product is the optimization layer that changes most frequently through iteration.
Learning Milestones Over Output Milestones
Traditional milestones measure output — features shipped, lines of code, hours worked. Learning milestones measure progress by what the team has demonstrably learned about customers, so a team that ships nothing but proves its assumption was wrong has still made real progress.
Lean Manufacturing Heritage
The methodology borrows directly from Taiichi Ohno and Shigeo Shingo: small batches, fast feedback, eliminating waste, and respecting workers' knowledge. The translation to startups is that knowledge work has its own waste — unvalidated specs, untested assumptions, half-finished features — that can be exposed by the same disciplines.
The Engine of Growth
A startup's mechanism for achieving sustainable customer growth. It is not a marketing campaign or a launch event; it is the closed feedback loop by which existing customers cause new customers to appear, and it must be deliberately designed, measured, and tuned.
- Entrepreneurial Management
- A management framework for organizations operating under extreme uncertainty — the alternative to MBA-style planning and "just ship it" chaos.
- Vision
- The unchanging end state the entrepreneur is trying to create.
- Strategy
- The plan to reach the vision (business model, customers, partners) — the layer that gets pivoted.
- Product
- The end result of the strategy; the layer that changes most frequently through optimization.
- Learning Milestone
- A measurable unit of progress demonstrating the team has learned something true about its business.
- Just Do It School
- Ries's label for the no-process, all-action approach to startups — dangerous because it confuses motion with progress.
- Engine of Growth
- The mechanism by which a startup achieves sustainable, repeatable customer acquisition.
Multiple choice
In Ries's Vision–Strategy–Product pyramid, which layer is the one that gets pivoted?
Spot the issue
A team measures progress by counting features shipped per sprint and hours logged per engineer. According to Chapter 1, what's the main problem with this?
Multiple choice
Ries describes a startup's engine of growth as which of the following?
True / False
The "just do it school" of startup management is safe because constant action eventually surfaces the right product.
Define
Ries gives a deliberately broad definition: "A startup is a human institution designed to create a new product or service under conditions of extreme uncertainty." This expands entrepreneurship beyond garage founders to include intrapreneurs inside large corporations, illustrated by Intuit's SnapTax team, which built a hit mobile tax product by behaving like a startup inside a Fortune 1000 company.
The Universal Definition of a Startup
"A human institution designed to create a new product or service under conditions of extreme uncertainty." Size, industry, sector, and funding model are irrelevant — what defines a startup is the condition of operating without a known answer for who the customer is or what they value.
Intrapreneurship
Innovation done by employees inside an existing organization. Ries argues that large companies must cultivate intrapreneurs to survive disruption and that they need the same Lean Startup tools as independent founders — extreme uncertainty doesn't get easier just because the parent company has a stable revenue line.
The SnapTax Case Study
A small team at Intuit built a mobile app letting users file taxes by photographing their W-2. The team launched first only to California — a small-batch release — and within three weeks of national launch had 350,000 downloads. The story shows Lean methods working inside a Fortune 1000 company, not just a garage.
Islands of Freedom
Scott Cook's term, cited by Ries, for the protected internal environments large companies must create so intrapreneurial teams can run experiments without being crushed by corporate process. Without these zones, every experiment must justify itself against the parent company's existing margin expectations and gets stalled.
Catalysts for Internal Startups
For an internal startup to function, Ries identifies three structural needs: scarce-but-secure resources (a small but uninterruptible budget), independent authority to develop the business, and a personal stake in the outcome. Without all three, the team reverts to executing on the parent's existing plans.
A Product Is Broadly Defined
In Ries's framework, a "product" includes any source of customer value — services, mobile apps, government programs, nonprofits, or physical goods. This matters because it means anyone delivering new value under uncertainty can use the methodology, not just software founders.
- Startup (Ries's Definition)
- A human institution designed to create a new product or service under conditions of extreme uncertainty.
- Extreme Uncertainty
- A condition in which the customer, market, and right product are all unknown; standard forecasting fails.
- Intrapreneur
- Someone driving entrepreneurial innovation from inside an established organization.
- Human Institution
- Ries's framing — a startup is fundamentally an organization of people, not just a product or technology.
- Island of Freedom
- A protected internal team or sandbox inside a large company where new ventures can experiment safely.
- Sustaining vs. Disruptive Innovation
- A Christensen distinction Ries uses to explain why large firms excel at incremental gains but struggle with truly new products.
Multiple choice
Ries defines a startup as:
Spot the issue
A VP at a Fortune 500 company hands an internal innovation team a generous budget and tells them to "go disrupt our industry," but every release must clear the same legal, brand, and revenue-impact reviews as the core product. Based on Chapter 2, what's missing?
Multiple choice
Which trio does Ries identify as the structural needs of an internal startup so that it can actually function?
Multiple choice
The SnapTax case study illustrates which point?
Learn
Ries reframes the meaning of progress: the only legitimate yardstick is validated learning — empirical demonstration that the team has discovered useful truths about its business. The IMVU story shows how shipping a product no one wanted produced "achieving failure" — flawless execution of a flawed plan — and how listening to customers turned the company around.
Validated Learning
The unit of progress for a startup: empirical demonstration, not after-the-fact storytelling, that the team has discovered actionable truths. The proof is data from experiments performed on real customers, and the test is whether the team would now make a different decision than before.
Achieving Failure
Successfully, faithfully, and rigorously executing a plan that turns out to have been wrong. Hitting every deadline on a product no one wants is not success. The danger is that the team's discipline and competence become evidence used to defend the wrong direction.
The IMVU Wrong-Bet Story
Ries's team spent six months building IM-network interoperability, certain customers would want to bring existing buddy lists. Customers didn't care — they wanted to meet new people in a standalone network. All that engineering was waste, and the lesson became the founding case study of the methodology.
Productivity Redefined
In a startup, productivity is not measured by features shipped or code written; it is measured by validated learning achieved per unit of time and money. A team that ships ten features and learns nothing is less productive than one that ships one feature and learns the entire approach is wrong.
The Question Behind Every Feature
Before building anything, ask: "If we succeeded at building this feature, would it matter?" If the answer isn't a clear yes backed by evidence about customer behavior, the work is potential waste no matter how well executed.
Two Kinds of "Learning"
The bad kind: vague, post-hoc rationalization used to excuse failure — "well, we learned a lot." The good kind: validated learning, supported by data from real customers, that would let the team predict what would happen if they tried the same thing again.
- Validated Learning
- Empirical demonstration that a team has discovered valuable, actionable truths about its business prospects.
- Achieving Failure
- Successfully executing a plan that does not lead to a viable business — failure disguised as success because milestones were "hit."
- Waste
- Any work that does not contribute to validated learning about customers.
- Hypothesis
- A testable, falsifiable prediction about customer behavior that drives an experiment.
- Vanity Metrics
- Numbers that look impressive (cumulative signups, total page views) but don't correlate with the drivers of a sustainable business.
- Actionable Metrics
- Numbers that demonstrate clear cause and effect, so a team can act on them and learn.
- Small Batches
- A lean-manufacturing principle that says working in tiny, frequent increments surfaces problems faster than large, infrequent releases.
Multiple choice
In Chapter 3, Ries says that the legitimate measure of startup progress is:
Spot the issue
A team hits every milestone on schedule and ships exactly the product they planned twelve months ago — but almost no one uses it. The CEO says, "At least we executed well." What's Ries's diagnosis?
Multiple choice
What does the IMVU wrong-bet story primarily illustrate?
Multiple choice
Before building any feature, Ries says the team should ask:
Experiment
Ries treats every feature, business plan element, and marketing strategy as a scientific experiment to be run, not a plan to be executed. The two most important hypotheses any startup must test are the value hypothesis (does this create real value?) and the growth hypothesis (how will customers discover it?), illustrated through Zappos, HP, and Village Laundry Service.
Think Big, Start Small
A startup's grand vision should be backed up by tiny, concrete experiments designed to test the leap-of-faith assumptions underneath it. The vision sets direction; the experiment limits exposure when the assumption turns out to be wrong, which it often does.
Value Hypothesis
A testable belief about whether a product actually delivers value to customers once they use it. The test is behavior — do they keep using it, recommend it, pay for it — not stated opinions in a survey, which are notoriously unreliable predictors of action.
Growth Hypothesis
A testable belief about how new customers will discover and adopt a product, and whether that mechanism is sustainable. A growth hypothesis that works once via a PR stunt is not validated; the test is whether the mechanism keeps producing new customers from past ones.
Leap-of-Faith Assumptions
The riskiest, most important beliefs underlying the entire business model. They must be identified and tested first, before optimizing anything else — there's no point tuning a conversion funnel for a value hypothesis that turns out to be false.
Genchi Gembutsu
Japanese for "go and see for yourself" — a Toyota principle Ries imports. Founders must talk to real customers in their real environments, not rely on reports. The example: Toyota engineer Yuji Yokoya drove ~53,000 miles across North America to design the 2004 Sienna minivan.
The Zappos MVP
Founder Nick Swinmurn tested whether people would buy shoes online by photographing inventory at local stores, listing them on a basic site, and buying them at retail when orders came in. A near-zero-cost experiment validated the value and growth hypotheses before any logistics infrastructure existed.
Concierge MVP
Manually delivering the product experience to a tiny number of customers, one at a time, to learn before automating anything. Village Laundry Service in India tested on-the-spot mobile laundry with a single truck and one customer at a time before scaling to kiosks.
- Value Hypothesis
- A testable statement of whether a product delivers value to customers when they use it.
- Growth Hypothesis
- A testable statement of how new customers will discover and adopt the product.
- Leap-of-Faith Assumption
- A foundational belief that, if wrong, invalidates the entire venture.
- Genchi Gembutsu
- Japanese for "go and see for yourself" — the Toyota practice of grounding decisions in firsthand observation.
- Concierge MVP
- A type of MVP where the team manually delivers the service to a tiny set of customers to validate demand before automating.
- Smoke Test
- An experiment (often a landing page or pre-order form) that tests whether customers will commit to a product that doesn't yet fully exist.
- A/B Test
- A controlled experiment comparing two product variants on comparable user groups to attribute behavior differences to the variant.
- Customer Archetype
- A concrete profile of the target customer used as a hypothesis to focus experiments.
Multiple choice
Ries identifies two hypotheses as the most critical to test in any startup. They are:
Spot the issue
A team launches a slick marketing stunt that drives a one-time burst of signups, then celebrates that their growth hypothesis is validated. What's wrong with this conclusion?
Multiple choice
Genchi Gembutsu — illustrated by Toyota engineer Yuji Yokoya driving ~53,000 miles across North America before redesigning the Sienna minivan — refers to:
Multiple choice
Nick Swinmurn wanted to know whether people would buy shoes online. Instead of building a warehouse, ordering inventory, and writing logistics software, he photographed shoes at local stores, posted them on a basic website, and bought from the stores at retail when orders came in. Which Lean Startup pattern best describes this?
Part 02
Steer — Build, Measure, Learn
Ch. 5–9
Leap
Every startup is built on a set of assumptions that must be true for the business to succeed; Ries calls the riskiest of these "leap-of-faith" assumptions. Before any building begins, founders should isolate their value and growth hypotheses, ground them in analogs and antilogs from prior companies, and validate by direct contact with real customers — not market reports or speculation.
Identifying the Leap-of-Faith
Not every business assumption is equally risky. Founders must rank assumptions by how much depends on each being true and how poorly understood each is. The top two are usually value and growth; everything else can wait until those are validated.
Analogs and Antilogs
Borrowed evidence from prior companies. An analog is "this worked for them, so it can work for us" (Apple's iPod used the Sony Walkman as analog). An antilog is a known counter-case that warns of risk (Napster was the iPod's antilog: would people still pay for music?). The gap between them defines what's still unknown.
Customer Archetype
A concise hypothesis-document describing the target customer's pain, motivations, and behaviors. Unlike a static persona, it is a testable artifact sharpened by real-world contact and discarded when the data says it's wrong.
Get Out of the Building
Steve Blank's mantra: founders cannot test assumptions from a conference room. Direct conversations with prospects expose what's actually unknown — and surface needs, objections, and workarounds that no survey or report would have captured.
Strategy vs. Vision Distinction
Vision is the destination — the world the founder wants to create. Strategy (hypotheses about customer, partners, growth) is how you reach it. Product is the most-frequently-changed output. Pivots change strategy without abandoning vision; persevering doesn't mean refusing to change strategy.
Why Surveys Lie
Customers reliably overstate their willingness to pay, use, or change behavior when asked in the abstract. The fix is observational evidence: what they do, not what they say. A signup form behind a smoke test produces sharper data than ten focus groups.
- Leap-of-Faith Assumption
- An unproven, business-critical belief that, if wrong, invalidates the venture.
- Value Hypothesis
- A testable prediction that customers will find the product valuable enough to use repeatedly.
- Growth Hypothesis
- A testable prediction of the mechanism by which the product attracts new users.
- Analog
- A successful precedent from another company or industry that supports an assumption.
- Antilog
- A cautionary precedent that highlights what is still unknown or risky.
- Customer Archetype
- A short profile of the target customer used as a hypothesis to be validated.
- Get Out of the Building
- Steve Blank's directive to test assumptions through firsthand customer contact.
Multiple choice
Among all the assumptions in a founder's business plan, which two does Ries say are usually the riskiest and most important to test first?
Multiple choice
A founder cites the Sony Walkman as evidence that a portable, personal music device can succeed, and cites Napster as a reminder that customers may resist paying for music. In Ries's vocabulary, what are these two references?
True / False
A well-designed customer survey, with a large enough sample, is the most reliable way to validate a leap-of-faith assumption about willingness to pay.
Spot the issue
A founding team has a strong vision and spends two months in a conference room writing a detailed customer archetype document based on industry reports and internal debate. They feel ready to start building. What's the main problem?
Test
Once leap-of-faith assumptions are identified, founders test them with a Minimum Viable Product — the smallest thing that produces validated learning. MVPs are not low-quality versions of the eventual product; they are deliberately stripped-down experiments (videos, concierge services, Wizard-of-Oz fakes) that risk small amounts of work to answer "should we build this at all?"
Minimum Viable Product (MVP)
The version of a new product that enables a full Build-Measure-Learn loop with the least effort. Its purpose is learning, not revenue or polished features. The right MVP is the one that lets a specific hypothesis fail informatively if it's going to fail.
Video MVP — Dropbox
Drew Houston's team made a 3-minute demo video showing seamless file sync across devices. With no working product, it drove beta signups from 5,000 to 75,000 overnight, validating demand before engineering a hard distributed system.
Concierge MVP — Food on the Table
The startup served one paying family at a time, with founders personally shopping, picking recipes, and producing grocery lists by hand. Each new family was added only after the previous experience was clearly working — and only later was the manual work automated into software.
Wizard of Oz MVP
Customers believe they're using a fully automated product, but humans perform the work behind the scenes. Aardvark's social-search Q&A was initially relayed by people; the team validated demand for the experience before building any of the automation it would need at scale.
Quality Defined by the Customer
Until you know what customers value, you cannot know what counts as a defect. IMVU shipped low-fidelity avatar "teleportation" because realistic motion was hard — and customers preferred the simple version. Internal engineering standards aren't a substitute for customer-defined quality.
MVP Risks and Mitigations
MVPs invite criticism because they ship imperfect work — risks include brand damage, competitor copying, and team morale. Ries recommends working under the radar or using a separate brand for risky experiments, and notes that competitor copying is rarely the real threat in practice.
- MVP (Minimum Viable Product)
- The smallest experiment that yields validated learning about a leap-of-faith assumption.
- Build-Measure-Learn Loop
- The iterative cycle of building an MVP, measuring response, and deciding to pivot or persevere.
- Concierge MVP
- A hand-delivered version of the service, one customer at a time.
- Wizard of Oz MVP
- A façade product whose "automation" is secretly performed manually.
- Video MVP
- A demonstration video used as an MVP, often to measure demand by signup rate.
- Smoke Test
- A pre-product signal that customers will commit (signups, pre-orders, click-throughs).
- Customer-Defined Quality
- Quality measured by what users actually value, not by internal engineering standards.
Multiple choice
What is the primary purpose of a Minimum Viable Product in Ries's framework?
Multiple choice
Drew Houston's Dropbox team used a three-minute demo video instead of a working product to drive beta signups from 5,000 to 75,000. Which MVP pattern does this illustrate?
Spot the issue
A team building a "smart" recipe-recommendation app secretly has employees email handcrafted recommendations to early users, while customers believe an algorithm is producing them. A senior engineer objects that this is dishonest and a waste of engineering capacity. What does Ries say?
True / False
Because an MVP ships with rough edges and missing features, a team that adheres to its internal engineering quality standards should refuse to release one.
Measure
A startup's job is to demonstrate with data that it is making progress toward a sustainable business — and traditional accounting cannot do this when the numbers start near zero. Ries introduces innovation accounting, a three-step measurement discipline backed by cohort analysis and split-testing, and warns against the seductive vanity metrics that hide the truth.
Innovation Accounting
A measurement framework that proves whether product changes actually improve customer behavior. It replaces revenue and headcount as progress signals with learning milestones suitable for early-stage products whose absolute numbers are still tiny.
Three Learning Milestones
(1) Establish a baseline with an MVP — concrete numbers for current behavior; (2) tune the engine by iterating toward the ideal model; (3) make the pivot-or-persevere decision when tuning stops yielding meaningful gains. Each milestone is gated by data, not opinion.
Vanity Metrics
Numbers that trend up and feel good but mask the truth: total registered users, raw pageviews, cumulative downloads, gross revenue without context. They confuse activity with progress — they almost always go up regardless of whether the underlying business is improving.
Actionable Metrics
Numbers tied directly to a specific action so causality can be inferred (e.g., conversion rate per cohort, retention curve per week). The test: can a team look at the number and decide what to do next? If not, it's probably vanity.
The Three A's
Good reports should be Actionable (clear cause-effect), Accessible (readable by anyone in the company, with people not just events), and Auditable (traceable back to source data so anyone can verify). Innovation accounting falls apart if any of the three is missing.
Cohort Analysis
Instead of looking at cumulative totals, group customers by the period they joined and watch how each cohort behaves over time. This reveals whether the product is genuinely improving — newer cohorts behaving better than older ones — rather than just stacking more users on a flat product.
Split-Testing (A/B)
Simultaneously offering two product variants to comparable groups so any difference in behavior is attributable to the variant, not seasonality or external noise. Without split tests, teams tend to credit shipped features for trends that would have happened anyway.
- Innovation Accounting
- Accounting designed for startups — tracks validated learning toward a sustainable business model.
- Vanity Metric
- A metric that grows even when the underlying business is failing.
- Actionable Metric
- A metric tied directly to a specific action so causality can be inferred.
- Cohort
- A group of customers who share a defining event (e.g., signed up the same week), tracked over time.
- Funnel Metrics
- Step-by-step customer behaviors (e.g., registration → activation → retention → referral → revenue).
- Baseline
- The initial data set from an MVP that establishes "where we are right now."
- The Three A's
- Actionable, Accessible, Auditable — the qualities of useful innovation-accounting reports.
Multiple choice
What are the three learning milestones of innovation accounting, in order?
Spot the issue
A startup's monthly board deck proudly shows that cumulative registered users have grown every single month for two years, and total page views are at an all-time high. The CEO concludes the product is succeeding. What's the main risk?
Multiple choice
Why does Ries argue cohort analysis is more honest than looking at cumulative totals?
Multiple choice
Ries says innovation-accounting reports should satisfy "the Three A's." Which trio is correct?
Pivot (or Persevere)
When innovation accounting shows the engine is not tuning toward the ideal, the team must pivot — a structured change of one element of the business hypothesis while keeping what has been learned. Ries argues most startups fail by pivoting too late and that scheduled pivot-or-persevere meetings prevent denial; a startup's true runway is the number of pivots it can still afford.
Pivot Definition
A structured course correction designed to test a new fundamental hypothesis about the product, strategy, or engine of growth — while preserving the vision and prior learning. Pivots are not wholesale do-overs; they replace one piece of the strategy and keep the rest.
Pivot-or-Persevere Meeting
A scheduled review (every few weeks to a few months) where product and business leaders together examine the data and explicitly decide to keep tuning or change a hypothesis. Forcing the meeting onto the calendar prevents the drift that lets failing strategies coast for months.
Runway as Pivots, Not Months
Runway is not just months of cash; it is the number of pivots the startup can still afford. Working in smaller batches and cheaper experiments doesn't just save money — it extends real runway by letting the team test more hypotheses before the cash runs out.
Failure to Pivot
Productive failure is rare; most startups die because they avoid the hard call. Vanity metrics, sunk-cost emotion, and "just-ship-one-more-feature" thinking all delay pivots — usually until there's no runway left to pivot with.
The Ten Pivot Types
Ries catalogs ten structured pivot patterns: zoom-in (one feature becomes the whole product), zoom-out (whole product becomes one feature), customer segment, customer need, platform (app↔platform), business architecture (high-margin/low-volume ↔ low-margin/high-volume), value capture (monetization model), engine of growth, channel, and technology.
Pivot ≠ Failure
Pivoting is the expected mode for early-stage companies and a sign of rigor, not weakness. The right question isn't "did we pivot?" but "did we learn enough between pivots to make this one a genuinely new hypothesis rather than a random guess?"
- Pivot
- A structured hypothesis change that preserves the vision.
- Persevere
- Keep iterating on the current strategy because metrics show progress.
- Pivot-or-Persevere Meeting
- A regular scheduled decision forum combining product and business leadership.
- Runway (Pivot Definition)
- The number of remaining pivots a startup can afford, not just calendar time.
- Zoom-In Pivot
- One feature of the previous product becomes the whole product.
- Zoom-Out Pivot
- The previous whole product becomes a single feature of a larger product.
- Customer Segment Pivot
- Same problem, different group of customers.
- Business Architecture Pivot
- Switch between high-margin/low-volume and low-margin/high-volume models.
Multiple choice
According to Ries, how should a startup actually measure its runway?
Multiple choice
A startup discovers that one minor feature of its product — a sharing tool buried inside a larger suite — is what users actually love and use. The team decides to discard the rest and rebuild the entire company around that one feature. Which pivot is this?
Spot the issue
A founder believes deeply in the original vision and has shipped twelve straight months of new features. Vanity metrics are slowly climbing, but cohort retention is flat. The founder argues "we just need to ship one more feature before considering a pivot." What does Ries diagnose?
True / False
A pivot represents a failure of the team's prior strategy and indicates the startup is in serious trouble.
Batch
The faster a startup can run the Build-Measure-Learn loop, the more learning it can buy with its runway — so batch size is a strategic weapon. Drawing on the Toyota Production System, Ries shows that smaller batches surface defects sooner, reduce work-in-progress waste, and enable practices like continuous deployment and the andon cord.
Small Batches Are Counterintuitively Faster
Doing one (or few) units of work end-to-end before starting the next seems slower than batching but is usually faster because it avoids hidden rework and queue time. The envelope-stuffing experiment from *Lean Thinking* — stuffing one envelope at a time beats fold-all/stamp-all/seal-all once setup, sorting, and error correction are counted — is the canonical illustration.
Work-in-Progress Is Waste
In a startup, WIP isn't physical goods; it's unvalidated specs, half-built features, and untested assumptions. WIP is pure waste until customers confirm it, because anything not validated may have to be thrown away when the next pivot lands.
Push vs. Pull Systems
Push systems make things based on forecasts (hope). Pull systems make things only when the next stage signals demand. Startups should pull features from validated customer demand, not push them from internal speculation or roadmap commitments.
Andon Cord
Toyota's line-stop mechanism: any worker can halt the entire production line on detecting a defect, forcing immediate root-cause fixing instead of letting defective units flow downstream. Works only in small-batch systems where stopping is cheap.
Continuous Deployment
The engineering practice of releasing code to production constantly. IMVU averaged ~50 deploys per day; each change is small, automatically tested, and instantly rolled back if cluster metrics regress. It is the software equivalent of single-piece flow.
Cluster Immune System
IMVU's automated guardrail: each push is exposed to a small slice of users; if revenue, registrations, or performance regress, the change is auto-reverted. This is what lets small batches ship safely without a long QA cycle.
The Large-Batch Death Spiral
When defects appear late, organizations respond by adding inspection, planning, and approval steps, which enlarge batches further, hiding defects even longer. The escape is to shrink batches enough that defects can be caught and fixed immediately.
- Batch Size
- Number of units processed at one step before moving to the next.
- Single-Piece Flow
- A batch size of one — the limiting case of small batches.
- Work-in-Progress (WIP)
- Partially completed work; in startups, includes unvalidated assumptions and untested features.
- Push System
- Production driven by forecasts or schedules upstream of demand.
- Pull System
- Production driven by actual downstream demand signals (Kanban).
- Andon Cord
- An any-worker stop signal that halts production when a defect appears.
- Continuous Deployment
- Continuously delivering code to production in tiny, automatically-tested increments.
- Cluster Immune System
- Automated production monitoring that rolls back deploys when key metrics degrade.
Multiple choice
In the envelope-stuffing experiment from *Lean Thinking*, doing one envelope end-to-end (fold, stamp, seal, address) before starting the next usually beats batching all the folds, then all the stamps, etc. Why does Ries cite this?
Multiple choice
In a startup, what counts as work-in-progress (WIP) that should be treated as waste until validated?
Spot the issue
A team's release process bundles three months of changes into a single quarterly deploy. When a defect ships, the response is to add two more rounds of pre-release inspection and a sign-off committee. Six months later defects are worse, not better. What pattern is this?
True / False
Toyota's andon cord — which lets any worker halt the entire production line on detecting a defect — only works well in systems that already use small batches.
Part 03
Accelerate — Scaling Lean Without Losing Speed
Ch. 10–14
Grow
Ries introduces sustainable growth — "new customers come from the actions of past customers" — and identifies three engines of growth (sticky, viral, paid) that power that feedback loop. Each engine has its own actionable metrics, its own logic, and its own saturation point; startups must focus on one engine at a time and know when its limits force a pivot.
The Sustainable Growth Rule
Long-term growth must come from the actions of past customers — repeat usage, word-of-mouth, paid advertising funded by past revenue, or side effects of usage. One-time bursts (PR stunts, launches) don't qualify as growth even when the chart spikes.
The Sticky Engine
Growth depends on a high retention rate: if the rate of new-customer acquisition exceeds the churn rate, the product grows. The key metric is the compounding rate (new minus churn), not gross signups — sticky businesses live or die on whether existing customers keep coming back.
The Viral Engine
Growth is a side effect of normal product use, tracked by the viral coefficient (new customers per existing customer). A coefficient above 1.0 yields exponential growth; below 1.0, growth dies out quickly. Hotmail's "PS: I love you" signature is the classic example.
The Paid Engine
Growth comes from spending part of the revenue per customer on acquiring the next one. Sustainable when Customer Lifetime Value (LTV) exceeds Cost Per Acquisition (CPA); the margin between them is the engine's marginal profit and the fuel available to scale.
One Engine at a Time
Each engine has different metrics, hypotheses, and operations, so a team that tries to drive growth on two engines simultaneously ends up running neither well. Focus on the dominant engine, tune it until it plateaus, then decide whether to add another.
Engine Tuning vs. Pivoting
Inside one engine, every experiment is a tuning move that should improve the engine's drivers (churn rate, viral coefficient, LTV/CPA). When tuning stops producing meaningful improvement, that's a signal to pivot to a new engine or strategy, not to push harder on the same lever.
Engines and Product/Market Fit
Rather than the vague "you'll know it when you feel it" definition, an engine of growth that is actually turning gives a quantitative way to recognize product/market fit: the loop is closed and accelerating on its own.
- Sustainable Growth
- Growth produced by past customers' actions, in a self-perpetuating loop.
- Sticky Engine
- A growth engine driven by high retention and a low churn rate.
- Viral Engine
- A growth engine where new users are a side effect of existing users' product use.
- Paid Engine
- A growth engine where part of revenue per customer funds acquisition of the next.
- Churn Rate
- The fraction of customers in a given period who stop engaging with the product.
- Viral Coefficient
- Average number of new users each existing user brings in; >1.0 means exponential growth.
- Customer Lifetime Value (LTV)
- Total revenue a customer contributes over their lifetime, minus variable costs.
- Cost Per Acquisition (CPA)
- Marginal cost of acquiring one new customer through paid channels.
Multiple choice
According to Ries's sustainable growth rule, new customers must come from where for growth to count as sustainable?
Spot the issue
A subscription SaaS team is simultaneously running aggressive paid acquisition, redesigning the product for virality, and rolling out referral incentives — all in the same quarter. What's the main risk in Lean Startup terms?
Multiple choice
For the paid engine of growth to be sustainable, which relationship must hold?
True / False
For a sticky engine of growth, the most important metric is the volume of gross new signups in the period.
Adapt
As startups scale, processes that worked at five people break under twenty; Chapter 11 shows how to build an adaptive organization that automatically adjusts its own discipline and speed. The central tool is the Five Whys — adapted from Toyota — which converts each failure into a proportional process investment, letting the team go faster *and* be more rigorous over time without freezing into bureaucracy.
Adaptive Organization
An organization that automatically adjusts its processes and performance to current conditions. Discipline is calibrated to actual problems surfaced by the team, not imposed top-down from a playbook of "best practices" that may not apply.
Just-in-Time Scalability
Build operational capacity (training, process, infrastructure) only when validated learning demands it, rather than over-engineering up front. Run experiments before you have permission to fail — and only invest in scale-out infrastructure when the engine of growth has been validated.
The Five Whys
A root-cause technique from Taiichi Ohno: ask "why?" five times in a row, drilling past symptoms to the underlying human or systemic cause. At the heart of every technical problem is a human problem — usually a missing training, a missing check, or a missing communication.
Proportional Investment
At each "why," make a small but meaningful corrective investment sized to the severity of the symptom. Small problems get small fixes; recurring or expensive problems get larger ones. This prevents the swing between under-investment (chaos) and over-investment (bureaucracy).
Automatic Speed Regulator
The Five Whys acts like a governor: when you accelerate too fast, more problems surface, which forces more process investments, which slows you to a sustainable pace; when things are calm, you can accelerate. The team finds its own right speed instead of having one imposed.
Five Blames Pitfall
Without discipline, Five Whys degenerates into finger-pointing — "Five Blames". Ries's rules: be tolerant of all mistakes the first time, never allow the same mistake twice, ensure everyone affected attends, and appoint a Five Whys master to moderate and own follow-up.
- Adaptive Organization
- A company whose processes self-tune to current conditions.
- Five Whys
- Iterative interrogation technique used to trace symptoms to root causes by asking "why" five times.
- Five Blames
- The dysfunctional form of Five Whys in which the conversation devolves into blaming individuals.
- Proportional Investment
- Sizing remedial process investments to match the cost of the defect that triggered them.
- Just-in-Time Scalability
- Adding operational capacity only when learning proves it's needed.
- Five Whys Master
- Designated moderator who runs the meeting and follows up on agreed actions.
- Taiichi Ohno
- The Toyota engineer who invented the Toyota Production System and the Five Whys method.
Multiple choice
Ries says that at the heart of every technical problem uncovered by the Five Whys is usually what?
Spot the issue
After a production outage, a team's "Five Whys" session quickly turns into a debate over which engineer pushed the bad code and whether they should be put on a performance plan. What's gone wrong?
Multiple choice
The principle of proportional investment in Five Whys means the team should...
True / False
A team should build out full operational infrastructure (training, scaling, formal processes) early so they are ready when growth arrives.
Innovate
Established companies want startup-style innovation but rarely create the conditions for it. Chapter 12 prescribes three structural attributes that any internal startup needs — scarce-but-secure resources, independent authority to build, and personal stake in the outcome — and introduces the innovation sandbox, a bounded zone where teams can experiment safely while the parent organization keeps running.
Three Structural Attributes for Internal Startups
Any internal startup needs (1) scarce but secure resources — a small but uninterruptible budget, because corporate politics is more dangerous than scarcity; (2) independent authority to develop its business — a cross-functional team that can ship without excessive approvals; (3) a personal stake in the outcome — credit, recognition, or career advancement tied to success.
The Innovation Sandbox
A bounded area of the product, customer base, or market within which an internal team is free to run experiments without parent-org approval. Successful inventions then graduate out of the sandbox into the main organization; the sandbox doesn't replace the existing business, it feeds it.
Seven Rules of the Sandbox
Experiments must (1) affect only sandboxed parts of the product or specific customer segments; (2) be owned end-to-end by one team; (3) have a fixed time limit; (4) affect a capped number/percentage of customers; (5) be evaluated on the same 5–10 actionable metrics; (6) use the same metrics across all teams in the sandbox; (7) be monitored in real time and aborted if something catastrophic occurs.
Entrepreneur as a Job Title
Ries argues "Entrepreneur" should be a recognized career path inside large companies, with its own job description, accountability via innovation accounting, and promotion criteria — not an honorific or a side project that everyone forgets about at performance-review time.
Transparent, Not Skunkworks
Unlike a hidden skunkworks lab, the sandbox is transparent: the parent organization knows what's running and on what metrics. This builds trust and prevents political backlash when innovations graduate, because nobody is surprised by the new product's existence.
Portfolio of Work
A company simultaneously runs four kinds of work: existing/legacy business, optimization/scaling, growth, and new innovation. Different leadership styles fit each phase, and talent should be allowed to migrate between phases as projects mature — not pinned to one mode forever.
- Innovation Sandbox
- Bounded environment within an established business where small experiments can run without endangering the larger organization.
- Internal Startup
- Entrepreneurial team operating inside a parent company under Lean Startup principles.
- Intrapreneur
- An internal corporate entrepreneur running an internal startup.
- Personal Stake
- The credit, recognition, or career upside an internal entrepreneur receives for successful outcomes.
- Scarce but Secure Resources
- A small, fixed budget that cannot be raided by other parts of the company.
- Cross-Functional Team
- A team with engineering, design, marketing, and other functions under one leader.
- Sandbox Graduation
- The process by which a proven experiment exits the sandbox and integrates into the main product.
- Portfolio of Work
- A management model balancing legacy operations, optimization, growth, and innovation simultaneously.
Multiple choice
Which trio does Ries identify as the three structural attributes that any internal startup needs in order to function?
Spot the issue
A Fortune 500 CEO funds an internal innovation lab as a secret skunkworks: nobody outside the lab knows what's running, on what metrics, or with which customers. Two years in, the first graduating product gets killed by political backlash from other divisions. What did the design get wrong?
Multiple choice
Per Ries's Seven Rules of the Sandbox, sandboxed experiments must be evaluated how?
Multiple choice
Ries's portfolio of work model says a company simultaneously runs which four kinds of work?
Epilogue: Waste Not
Ries closes by locating Lean Startup in a longer intellectual lineage that begins with Frederick Winslow Taylor and scientific management. Taylor revolutionized industry by insisting that work could be studied and improved, but his "science" was misread as rigid technique, producing exactly the bureaucratic dysfunction Lean Startup is designed to fight. In the 21st century the real waste is failure to learn — building things customers don't want.
Scientific Management Lineage
Taylor's *Principles of Scientific Management* (1911) is the conceptual ancestor of Lean manufacturing, the Toyota Production System, and Lean Startup. Taylor effectively invented "management" itself and the idea that work can be studied and improved through conscious effort — not just done.
Science vs. Rigid Technique
Taylor "preached science as a way of thinking," but his ideas were widely confused with the specific stopwatch-and-checklist techniques he advocated. Much of 20th-century industrial dysfunction came from importing the techniques without the scientific mindset — the same risk Lean Startup faces today.
A New Definition of Waste
In manufacturing, waste = physical effort that doesn't add product value. In a startup, waste = effort that doesn't produce validated learning. As Drucker put it, "there is surely nothing quite so useless as doing with great efficiency what should not be done at all."
Functional Efficiency vs. Learning
Departments optimizing their own throughput — engineers shipping features, marketers running campaigns — can look extremely efficient while the company as a whole is wasting effort because it's building the wrong thing. Lean Startup measures progress by learning, not by departmental output.
Taylorism as Cautionary Tale
Lean Startup is to knowledge work what Taylorism was to manual work: the application of scientific method to a domain previously thought to be pure art. But it inherits Taylor's failure mode — risk of being misread as rigid prescription — and Ries warns explicitly against turning it into dogma.
Future of Management
The next great managerial revolution will be applying scientific thinking to innovation itself, so that "startup success can be engineered by following the right process … which means it can be learned." This is the book's largest claim: entrepreneurship is teachable.
- Scientific Management
- Frederick Winslow Taylor's discipline of studying and improving work using systematic measurement.
- Frederick Winslow Taylor
- American mechanical engineer (1856–1915), author of *The Principles of Scientific Management*, conceptual forefather of Lean.
- Taylorism
- Common name for Taylor's school of management; in popular usage, often the rigid, dehumanizing version Ries cautions against.
- Waste (Lean Startup Definition)
- Any activity that does not contribute to validated learning about customers.
- Functional Efficiency
- Optimizing the output of a single department, often at the expense of overall company learning.
- Pseudoscience of Management
- Management practices that adopt the vocabulary of science without falsifiable hypotheses and honest measurement.
- Drucker's Dictum
- "There is nothing quite so useless as doing with great efficiency that which should not be done at all."
Multiple choice
In Ries's redefinition for the startup era, waste is best defined as...
Spot the issue
An engineering org celebrates a record quarter: features-per-sprint is up 40 percent and the marketing team has the highest campaign-launch rate in company history. Meanwhile, the product has shipped no validated learning about whether customers actually want any of it. What does Ries call this pattern?
Multiple choice
Ries traces Lean Startup's intellectual lineage to which historical figure and discipline?
True / False
According to Ries, the main lesson from Taylorism is that we should adopt its specific stopwatch-and-checklist techniques wholesale and apply them rigorously to knowledge work.
Join the Movement
The closing section is a call to action: Lean Startup has become a global movement, and the reader is invited to join it, contribute to it, and help develop its still-unfinished body of knowledge. Ries points to community resources — meetups, the Lean Startup Circle, ongoing experiments inside companies of every size — and frames Lean Startup not as a finished theory but as an open, collaborative practice.
A Movement, Not a Manual
Lean Startup is an evolving body of practice rather than a fixed methodology; the book is a starting point, not a final answer. New case studies, failures, and adaptations from practitioners are expected to keep changing the methodology over time.
Community of Practice
Practitioners refine the methodology by sharing experiments, failures, and adaptations across industries and company sizes. The methodology gets better as the community shares — and gets worse if it ossifies into a fixed set of techniques to be applied without thought.
Universal Applicability
Lean Startup is not just for tech founders: nonprofits, government agencies, schools, and large enterprises (GE, Toyota, Intuit) all face conditions of extreme uncertainty and can apply the method. The defining condition is uncertainty about customers, product, or business model — not industry.
Stop Wasting Time
The book's final exhortation: stop wasting time and resources building things customers don't want. Start running experiments, sharing what you learn, and demanding the same from teammates and colleagues — the cost of waste is now visible and avoidable.
Collective Responsibility
Everyone on a team — not just founders — shares responsibility for learning, adapting, and serving customers. The movement spreads horizontally, not just from CEOs down; an engineer or designer can run experiments inside a sandbox just as effectively as a founder.
Open Methodology
A body of practice that improves as its users contribute case studies, critiques, and refinements, rather than being closed and proprietary. The contrast is with management fads sold by consultants — Lean Startup is meant to be common property.
- The Lean Startup Movement
- The global community of practitioners applying and extending Eric Ries's methodology.
- Lean Startup Circle
- The primary online community (forum, mailing list, and meetup network) for practitioners.
- Meetup Group
- A local in-person gathering where practitioners share experiences and case studies.
- Community of Practice
- A group of practitioners who deepen expertise in a shared craft by exchanging knowledge.
- Practitioner
- Anyone — founder, intrapreneur, manager, designer — actively applying Lean Startup methods.
- Open Methodology
- A body of practice that improves through community contribution rather than being closed and proprietary.
- Call to Action
- Ries's explicit invitation to readers to apply, contribute to, and spread Lean Startup methods.
- Universal Applicability
- The claim that Lean Startup principles apply anywhere an institution faces extreme uncertainty.
Multiple choice
How does Ries frame Lean Startup at the end of the book?
Spot the issue
A government agency director dismisses Lean Startup, saying "this is fine for tech founders in Silicon Valley but it has nothing to do with how a public-sector agency runs." How would Ries push back?
True / False
Ries argues that only founders and CEOs are responsible for applying Lean Startup methods inside an organization.
Key Takeaways
A startup is a human institution creating something new under extreme uncertainty — it needs its own management discipline, not luck or the playbook of an established company.
Progress is measured in validated learning, not features shipped — perfectly executing a plan no customer wants is "achieving failure."
Every business plan rests on leap-of-faith assumptions — usually a value hypothesis and a growth hypothesis — that must be tested with cheap experiments, not predicted from a spreadsheet.
Innovation accounting (baseline → tune the engine → pivot or persevere) replaces vanity metrics with cohort analysis and split tests so teams know whether they're actually making progress.
Real runway is the number of pivots a startup can still afford, so shrink batch sizes to run more Build-Measure-Learn loops before the money runs out.
Lean Startup scales through adaptive practices — Five Whys, andon cord, innovation sandboxes — that let large organizations innovate without crushing the teams doing it.