📖 Eric Ries and the Origins of Lean Startup
In 2004, Eric Ries was CTO of IMVU, a startup for 3D avatars for chat. The team spent 6 months building a product nobody wanted. Months of code, design, and planning went to waste. This traumatic experience became the founding moment of Lean Startup: what if there were a method to find out what not to build before spending time and money?
🎯 The Central Thesis
Startups don’t fail due to a lack of product. They fail due to a lack of customers. The biggest risk isn’t technical; it’s market-related. So, a startup’s most important job isn’t to build; it’s to learn. And every activity that does not produce validated learning is waste.
Eric Ries, “The Lean Startup” (2011)
🧬 Lean Startup DNA
Lean Startup combines 3 influences:
- • Customer Development (Steve Blank): get out of the office and talk to customers before writing code
- • Lean Manufacturing (Toyota): eliminate waste, maintain continuous flow, and improve iteratively
- • Agile Development: short cycles, frequent delivery, rapid adaptation
🔄 The Build-Measure-Learn Cycle
The engine of Lean Startup is a feedback loop: Build → Measure → Learn. The goal is not to build the product faster. It is go through the entire cycle as quickly as possible. Each cycle tests a specific hypothesis and produces validated learning.
⚠️ The Counterintuitive Trick
You plan backward. First, decide what you want to learn (hypothesis). Then define how to measure it (metric). Only then decide what to build (MVP). Most teams do the opposite: build first, then try to measure something, and rarely learn anything useful.
// Wrong order (90% of teams):
Build → “How did it go?” → “Hmm, I don’t know” → Build again
// Correct order (Lean):
Learn (hypothesis) → Measure (metric) → Build (minimum MVP)
🧪 MVP: Types and Strategies
MVP (Minimum Viable Product) is the simplest product that allows test a hypothesis with real customers. It is not a “half-baked” version of the final product. It is an experiment. The MVP answers ONE question: is this worth building?
Wizard of Oz MVP
It seems automatic, but there are people behind it
The user sees a functional product, but a human does the work manually behind the scenes. Tests whether people want the functionality BEFORE spending months automating it. Example: a chatbot that is actually a human responding; a recommendation app where suggestions are curated manually.
Concierge MVP
Manual, personalized service for a few
Unlike the Wizard of Oz, the customer KNOWS it’s manual. You deliver the service one-on-one to learn what works before scaling. Example: Zappos (the founder bought shoes at local stores and shipped them manually); personalized diet consulting before creating an app.
Smoke Test / Landing Page MVP
Test demand before building anything
A page that describes the product and has a “I want it!” button or email form. It measures whether people are interested enough to click or sign up. No product code. Example: Buffer launched a landing page describing the service and collected emails before writing the first line of code.
Video MVP
Show the concept in a video and measure the reaction
A 3-minute video explaining what the product does (without the product existing). It measures engagement, shares, and sign-ups. The classic case: Dropbox created a demo video that generated 70,000 waitlist sign-ups overnight. Perfect validation.
💡 The MVP Golden Rule
If you’re not embarrassed by your MVP, you launched too late. The MVP isn’t about the product. It’s about the speed of learning. Every extra day spent “polishing” is a day without real feedback.
📊 Vanity Metrics vs. Actionable Metrics
The biggest Lean Startup trap isn't building the wrong MVP. It's measure the wrong thing. Vanity metrics make you feel good. Actionable metrics help you make better decisions. That difference is the difference between success and self-deception.
✗ Vanity Metrics
- ✗ Total registered users (includes dropouts)
- ✗ Total page views (one user in the loop = lots of views)
- ✗ Downloads (downloaded != used != paid for)
- ✗ "Followers" (not necessarily customers)
- ✗ Cumulative revenue (only goes up, never indicates a trend)
✓ Actionable Metrics
- ✓ Activation rate (% who complete the key action)
- ✓ Retention by cohort (do users from week X come back in week Y?)
- ✓ NPS (Net Promoter Score) (would you recommend it?)
- ✓ LTV/CAC (revenue per customer / acquisition cost)
- ✓ Revenue by cohort (real week-over-week trend)
🧮 Innovation Accounting
Ries calls the 3-step framework for measuring progress in uncertain environments Innovation Accounting:
- 1. Baseline: Launch the MVP and measure where you are today (actual conversion rate, not desired rate)
- 2. Tuning: Run experiments to move the metric from the baseline toward the ideal
- 3. Decision: If the experiments don’t move the metric, it’s time to pivot. If they do, persevere
🔀 Pivot vs. Persevere
The pivot is the hardest decision in Lean Startup. It means change strategy without changing vision. It is not giving up. It is redirecting based on evidence. Most successful startups pivoted at least once: Instagram started as Burbn (a check-in app), Twitter was Odeo (a podcasting platform), and Slack was a game called Glitch.
🔄 Types of Pivot
Zoom-in Pivot
A single feature becomes the entire product. Example: Instagram removed everything from Burbn except photos+filters.
Zoom-out Pivot
The entire product becomes a feature of something larger. Example: Yelp started as an email recommendation service among friends.
Customer Segment Pivot
Same product, different audience. E.g., Slack from gamers to businesses.
Channel Pivot
Same product, different distribution channel. E.g., from direct sales to a marketplace.
Value Capture Pivot
Change how you monetize. E.g., from a one-time sale to a subscription (Adobe Creative Cloud).
Platform Pivot
From application to platform (or vice versa). Example: Amazon from store to AWS+marketplace.
⚠️ Signs It’s Time to Pivot
- • Experiments aren’t moving the metrics after 3+ cycles
- • Customers use the product in unexpected ways (a sign of a pivot)
- • The team is "optimizing" without affecting user behavior
- • Conversations with customers reveal that the real problem is something else
📚 Classic Lean Startup Cases
The theory comes to life in the examples. These are the two most studied Lean Startup cases, and each teaches something different about validation.
📦 Dropbox: The Perfect MVP Video
Drew Houston had a problem: forgetting his USB drive. The technical solution (cloud file syncing) was complex. Investors were skeptical: "there are already 20 companies doing this." Houston didn't build the product. He made a 3-minute video showing how it would work.
Hypothesis: People want file sync if it's transparent (without thinking).
MVP: A 3-minute screencast video posted on Hacker News.
Metric: Sign-ups on the waitlist.
Result: 5,000 → 75,000 subscribers in 24h. Hypothesis validated.
👟 Zappos: The Original Concierge MVP
Nick Swinmurn wanted to sell shoes online in 1999. No one believed people would buy shoes without trying them on. Instead of setting up inventory and logistics, he went to local shoe stores, took photos, posted them online, and when someone bought a pair, he went to the store, bought it, and mailed it.
✓ What the MVP Validated
- • People buy shoes online (demand)
- • Willingness to wait for delivery
- • Photos are enough (no hands-on testing)
- • Return policy addresses the objection
📊 MVP Investment
- • Cost: almost zero (camera + basic website)
- • Time: days, not months
- • Risk: only bought when it was on sale
- • Result: acquired by Amazon for $1.2B
💡 Lean Startup in the Corporate World
Lean Startup isn’t just for startups. GE, Intuit, Toyota, and Bosch use the BML cycle internally. The difference: in corporations, the "pivot" is more political than technical. Innovation Accounting is essential for justifying learning (not just revenue) to the board.
📋 Module Summary
Next Learning Path:
Track 3 - Implementation: how to create corporate innovation programs