📖 Chapter 4: Experiment
The Lean Startup – Turning Ideas into Scientific Experiments
Overview: Why Experimentation Matters
In Chapter 3, we learned about Validated Learning. Now in Chapter 4, we dive deeper into HOW to structure that learning through systematic experiments.
The key insight: Every startup is essentially a grand experiment – an attempt to answer the question: “Can we build a sustainable business around this vision?”
What is a Startup Experiment?
A startup experiment is not just “trying something and seeing what happens.”
It’s a structured test where you:
- State a clear hypothesis about your customers
- Design a minimum test to validate or invalidate it
- Collect real data from real customer behavior
- Make a decision based on what you learned
The Two Most Important Hypotheses
Eric Ries explains that every startup should test two fundamental assumptions first:
1. The Value Hypothesis
Question: Does your product actually deliver value to customers?
Example: “Users will find our meal planning app valuable enough to use it daily.”
How to test: Build a simple MVP and measure actual usage (not just signups!)
2. The Growth Hypothesis
Question: How will new customers discover your product?
Example: “Users will refer their friends after trying our app.”
How to test: Track how your first 100 users found you and how many referred others.
☕ Hamed’s Analysis: Why These Two Come First
I always tell my clients: You need to prove these two things before anything else!
Why? Because:
- If you can’t deliver value → No one will keep using your product
- If you can’t grow → You’ll run out of customers
Real example: I worked with a client building a meditation app. We spent 2 weeks testing the Value Hypothesis first:
- Built a simple prototype with just 3 guided meditations
- Gave it to 50 people for free
- Tracked: Did they use it more than once?
Result: Only 8 out of 50 used it twice. This told us the VALUE wasn’t there yet – we needed to improve before worrying about growth!
Case Study: Village Laundry Service
One of the most powerful examples in this chapter comes from Village Laundry Service – a startup that offers laundry pickup and delivery.
The Initial Hypothesis
The founders believed: “Busy professionals will pay for convenient laundry pickup service.”
But they didn’t know if this was true!
The Experiment
Instead of building an app, hiring drivers, and buying equipment, they did something simple:
- Created a basic landing page explaining the service
- Put their personal phone number on it
- Ran a small Facebook ad targeting local professionals
- When people called, they personally picked up and delivered laundry!
Cost: About $200 for ads + their time
The Results
Within 2 weeks, they had 40 customers who were willing to pay!
Key learning: Yes, the Value Hypothesis was validated – people DO want this service.
Next step: Now they could confidently invest in building proper systems.
☕ Hamed’s Analysis: The Power of “Doing Things That Don’t Scale”
This is one of my favorite startup principles: “Do things that don’t scale!”
The Village Laundry founders could have spent 6 months building:
- A mobile app for customers
- A driver scheduling system
- Payment processing integration
- Automated notifications
But instead, they did it ALL manually first!
Lesson: Don’t automate until you’ve proven people want it manually!
My example – Restaurant Website:
A restaurant owner wanted online ordering on their website. Instead of building complex software, we did this:
- Week 1: Added a WhatsApp button to their existing website
- Week 2: Manually took orders via WhatsApp chat
- Week 3: Tracked how many people ordered vs. just browsed the menu
Result: 60% of people who clicked WhatsApp actually placed orders! NOW we knew it was worth building proper online ordering.
How to Design Effective Experiments
Eric Ries provides a framework for designing experiments that actually teach you something:
The 5-Step Experiment Design Process
Step 1: Write Down Your Assumption
Be specific! “Users want healthy food” is too vague. Better: “Office workers aged 25-40 will order healthy lunches if delivered within 30 minutes.”Step 2: Identify What Data Would Prove/Disprove It
What specific number or behavior would tell you if you’re right? Example: “If 30% of people who see our offer actually order, we’re onto something.”Step 3: Design the Minimum Test
What’s the simplest, fastest way to collect that data? Often it’s way simpler than you think!Step 4: Determine Success Criteria in Advance
Decide BEFORE the experiment: “If we get X result, we continue. If we get Y result, we pivot.”Step 5: Run the Experiment and Learn
Actually do it, collect real data, and make a decision based on what you learned.
“The goal of a startup is to figure out the right thing to build – the thing customers want and will pay for – as quickly as possible.”
(The goal isn’t building – it’s learning what to build!)
End of Part 1
In Part 2 we’ll cover:
• More real-world experiment examples
• How to avoid common experimentation mistakes
• Five Key Takeaways from Chapter 4
📖 Chapter 4: Experiment – Part 2
More Real-World Experiment Examples
Let’s look at more examples of how successful companies used experiments to validate their ideas:
Example 1: Dropbox’s Video MVP
The Challenge: Dropbox needed to test if people wanted file syncing software, but building it would take months.
The Experiment: Instead of building the product, founder Drew Houston made a 3-minute demo video showing how it would work.
The Result: The video was posted on a tech forum and overnight, their beta waiting list went from 5,000 to 75,000 people!
Key Learning: This validated the Value Hypothesis WITHOUT writing complex code.
Example 2: Zappos’ First Experiment
The Hypothesis: “People will buy shoes online without trying them on first.”
The Experiment: Founder Nick Swinmurn didn’t buy inventory. Instead, he:
- Took photos of shoes in local shoe stores
- Posted them on a simple website
- When someone ordered, he’d go buy the shoes from the store and ship them!
The Result: People DID buy shoes online! This validated the core assumption before investing in warehouses and inventory.
☕ Hamed’s Analysis: What These Examples Teach Us
Notice a pattern? None of these founders built the “real” product first!
They all found creative ways to test their core assumption with minimal investment:
- Dropbox → Made a video (cost: $0, time: 1 weekend)
- Zappos → Manual fulfillment (cost: minimal, time: immediate)
- Village Laundry → Personal pickup (cost: gas money, time: their evenings)
Key principle: Test the RISKIEST assumption first with the CHEAPEST method!
My example – Fitness Coach App:
I consulted for a fitness coach who wanted an app for workout plans. Instead of building an app, we tested this way:
- Week 1: Created a Google Form for workout requests
- Week 2: Sent workout plans as PDF files via email
- Week 3: Asked: “Would you pay $10/month for this?”
Result: 40 out of 50 people said YES! Only THEN did we start building the app.
Common Experimentation Mistakes to Avoid
Eric Ries warns against several common pitfalls when running startup experiments:
❌ Mistake #1: Analysis Paralysis
The Problem: Spending months planning the “perfect” experiment instead of running an imperfect one.
The Fix: Done is better than perfect! Run a quick, imperfect test this week rather than a perfect one in 3 months.
❌ Mistake #2: Vanity Metrics
The Problem: Measuring things that make you feel good but don’t indicate real progress.
Examples of Vanity Metrics:
- Total registered users (who never come back)
- Page views (from people who immediately leave)
- Social media followers (who don’t buy)
The Fix: Focus on actionable metrics like retention rate, customer lifetime value, and conversion rates.
❌ Mistake #3: Ignoring Negative Results
The Problem: Only paying attention to data that confirms what you already believe.
The Fix: Negative results are GOOD! They save you from wasting months on the wrong path. Embrace them!
❌ Mistake #4: Building Instead of Testing
The Problem: Building features before validating if anyone wants them.
The Fix: Always ask: “What’s the SMALLEST thing I can do to test this assumption?”
☕ Hamed’s Analysis: The Hardest Lesson
In my experience, the hardest mistake to avoid is #3 – Ignoring Negative Results.
Why? Because we get emotionally attached to our ideas!
Real story: I worked on a project for an online course platform. We believed: “People want live interactive courses more than pre-recorded ones.”
We tested it and guess what? Only 20% of users showed up for live sessions, but 80% watched recordings later!
The temptation: “Maybe we just need better marketing!” or “The timing was bad!”
The reality: Our assumption was WRONG. People wanted flexibility, not live interaction.
Lesson: When data contradicts your belief, believe the data!
The Scientific Method for Startups
Eric Ries emphasizes that Lean Startup applies the scientific method to business:
Traditional Science vs. Startup Experiments
Traditional Science:
- Form hypothesis
- Design controlled experiment
- Collect data
- Accept or reject hypothesis
- Publish results
Lean Startup Science:
- Form hypothesis about customers
- Build minimum viable test
- Collect data from REAL customers
- Validated learning: What did we learn?
- Decide: Persevere or Pivot
“Experiments are more than just theoretical inquiries. They are the first products. If this or any other experiment is successful, it allows the manager to get started with his or her campaign: enlisting early adopters, adding employees to each further experiment, and eventually starting to build a product. By the time that product is ready to be distributed widely, it will already have established customers. It will have solved real problems and will offer detailed specifications for what needs to be built.”
(Your experiments ARE your first products!)
Quick Action Plan: Design Your First Experiment
Here’s a practical template you can use RIGHT NOW:
Your Experiment Design Template
1. My Riskiest Assumption:
(Example: “Busy parents will pay $20/month for meal planning service”)2. How I’ll Test It:
(Example: “Create a landing page, run Facebook ads, collect emails of interested people”)3. Success Metric:
(Example: “If 15% of page visitors give their email, the assumption is validated”)4. Timeline:
(Example: “Run for 2 weeks with $200 ad budget”)5. What I’ll Do With Results:
(Example: “If validated → Build MVP. If not → Talk to 10 people who visited but didn’t sign up”)
☕ Hamed’s Analysis: Start THIS WEEK
Don’t wait for the “perfect” experiment design. Here’s what I tell every client:
Pick ONE assumption and test it THIS WEEK with whatever you have!
Examples of experiments you can run this week:
- Post about your idea on social media and count how many people comment asking for more info
- Create a Google Form describing your product and ask: “Would you pay $X for this?”
- Call 10 potential customers and describe your idea – count how many say “That’s interesting!”
- Make a simple landing page on Carrd (free!) and share it in relevant online communities
Remember: An imperfect experiment THIS WEEK is worth more than a perfect one in 3 months!
Five Key Takeaways from Chapter 4
✅ Key Takeaway #1: Every Startup is an Experiment
Your startup is not just a business – it’s a scientific experiment to test if you can build something customers want and will pay for.
✅ Key Takeaway #2: Test Value BEFORE Growth
First prove your product delivers value (Value Hypothesis), THEN worry about how to scale it (Growth Hypothesis). Don’t do it backwards!
✅ Key Takeaway #3: Do Things That Don’t Scale
Manual processes are GOOD in the beginning! They let you test assumptions quickly without building complex systems. Automate later!
✅ Key Takeaway #4: Design Before You Build
Before running an experiment, write down: (1) Your assumption, (2) How you’ll test it, (3) What success looks like, (4) What you’ll do with the results.
✅ Key Takeaway #5: Embrace Negative Results
Negative results are NOT failures – they’re valuable learning! They save you from wasting months on the wrong path. Celebrate them!
🎉 Chapter 4 Complete!
You now understand how to turn your startup ideas into scientific experiments!
Next up: Chapter 5 – Leap
(How to identify and test your leap-of-faith assumptions)
Leave a Reply