One loop. Eight sections. Zero theory. By the end of this — you'll never build blind again.
⚡ Core Principle → ASSUMPTION → TEST → RESULT → LEARN → DECISION → REPEAT
Section 01
THE ONE LOOP
This is the entire Lean Startup method. Everything connects back to this. Memorize it.
00
IDEA
Something you think could work
01
ASSUMPTION
What you believe to be true
02
TEST (MVP)
Smallest possible check
03
REAL USER RESPONSE
What actually happened
04
RESULT
Data. Not feelings.
05
LEARN
What the result actually means
06
DECISION
Stay, pivot, or kill
07
NEXT ASSUMPTION
New question. New test.
↺ REPEAT — always
The rule:
You never prove you're right. You just keep proving you're less wrong. Every loop gets you closer to what actually works.
Section 02
HOW TO THINK
This is the operating system inside your head. Run this on every idea, every decision, every week.
1. "What do I believe is true?"
Name the assumption. Be specific. "People will pay" is not specific. "HVAC owners will pay $497/mo to stop missing calls" is.
2. "How do I test this as fast as possible?"
What's the cheapest, fastest way to get a real answer? Not a survey. A test with skin in the game.
3. "What actually happened?"
Not what you hoped. What the numbers, replies, and behavior actually showed.
4. "What does that mean?"
Was the assumption right, partially right, or dead wrong? No spin. Just honesty.
5. "What do I do next?"
Change something, keep going, or stop. Then start the loop again.
▸ Service Business Example
OppScale / HVAC Cold Outreach
Believe: HVAC owners lose 4–6 bookings/week from missed calls and will pay to fix it.
Test: Cold call 30 HVAC owners this week. Lead with the loss, not the product.
Happened: 6 booked demos. 2 said "I didn't know that was even possible."
Means: The pain is real. The framing works. The pricing hasn't been tested yet.
Next: Get to the pricing conversation on the next 30 calls.
▸ SaaS / AI Example
AI Scheduling Tool for Clinics
Believe: Dental offices will pay for an AI that handles appointment confirmations automatically.
Test: Offer to set it up free for 3 clinics in exchange for a case study.
Happened: 2 of 3 loved it. But they kept turning it off — staff didn't trust the AI.
Means: The problem is real but adoption depends on staff trust, not just owner buy-in.
Next: Test with an onboarding session specifically for front-desk staff.
Section 03
BUILD → MEASURE → LEARN
This is just the loop with different labels. Here's what each word actually means in practice.
BUILD
Run the test
Make the call. Send the email. Post the landing page. Do the thing that generates a real response.
build the full product
MEASURE
Watch what people DO
Did they click? Did they buy? Did they reply? Did they ghost? Actions beat opinions every time.
ask what they think
LEARN
Decide what's next
Confirmed? Push harder. Failed? Change something. Unclear? Run a tighter test.
celebrate or spiral
Do I use this the whole time? → YES. Every phase of business. Every week. Forever. Markets change, assumptions expire, and what worked last quarter might not work now. The loop never ends — only the assumptions change.
Section 04
VALIDATING ASSUMPTIONS
Every test returns one of three results: Fail, Partial, or Win. Here's how to read each one.
ASSUMPTION → TEST → RESULT → DECISION
✗ FAIL — Wrong Assumption
Assumption
HVAC owners will respond to cold emails about missed calls
Kill the email channel. Switch to cold calls immediately.
~ PARTIAL — Half Right
Assumption
HVAC owners will pay $750 setup + $497/mo for AI receptionist
Test
Quote this exact price on 10 demo calls
Result
3 said yes to $497/mo but pushed back on setup fee
Decision
Test $0 setup + higher monthly. The recurring is the right play.
✓ SUCCESS — Confirmed
Assumption
Leading with "you're losing 4–6 bookings/week" will book demos
Test
Cold call 30 HVAC owners using this exact opener
Result
7 agreed to a demo. Previous script got 1 out of 30.
Decision
Lock this opener. Train it. Scale calls. Test close rate next.
Section 05
WHEN TO PIVOT
Most people pivot too early because they're scared. Or too late because they're attached. Here's a clean system.
IF
No one cares about the problem
→
CHANGE THE PROBLEM
IF
People care but won't pay / commit
→
FIX THE OFFER
IF
People pay and want more
→
KEEP GOING
▸ Example 1 — Change the Problem
AI Chatbot for Restaurants
Built a chatbot to handle reservations. Nobody cared — restaurants used OpenTable. The problem wasn't reservation management, it was no-shows. Pivoted to an AI that sends reminder texts and reduces no-shows by 40%. Signed 5 clients in 2 weeks.
▸ Example 2 — Fix the Offer
OppScale Pricing Test
HVAC owners loved the AI receptionist demo. But when quoted $750 setup + $497/mo, 6 of 8 said "let me think about it" and ghosted. The problem wasn't the product — it was upfront risk. Offer changed to $0 setup + 30-day trial. Close rate jumped to 4 out of the next 8.
Section 06
FULL WALKTHROUGH
OppScale Real Example
One idea. One real execution. Every step of the loop in action.
01
IDEA
HVAC companies miss calls when they're on jobs. Those missed calls = lost revenue. An AI voice receptionist could answer, qualify, and book automatically.
02
ASSUMPTION
HVAC owners in the $1M–$5M range are losing 4–6 bookings per week from missed calls, and they will pay to stop that loss.
03
TEST (MVP)
Cold call 30 HVAC owners from an Apollo list. Use the opener: "I work with HVAC companies that are losing 4–6 bookings a week to missed calls — is that happening for you?" Demo a fake company persona (SteelCity Air) if they want to see it.
04
RESULT
7 out of 30 booked a demo. 12 said "yes that's happening" but didn't book. 11 hung up or said not interested. Previous script: 1 demo per 30 calls.
05
LEARN
The pain is confirmed. The loss-framing opener works. The 12 who acknowledged the problem but didn't book = conversion problem, not a pain problem. Pricing hasn't been tested yet.
06
DECISION
Lock the opener. Run another 30 calls. This time get to pricing on every demo call. Also test a follow-up sequence for the 12 who said "yes but not now."
07
NEXT STEP
New assumption: "HVAC owners who book a demo will pay $497/mo with no setup fee." New test: Run 7 demos. Present pricing. Count how many say yes without stalling.
Section 07
COMMON MISTAKES
These kill businesses before they have a chance to find out if they work.
🏗️
Building Too Much
You built a full product before a single customer confirmed they'd pay — now you've wasted months on something nobody wants.
🕳️
Measuring Nothing
If you're not tracking what people actually do (clicks, replies, buys, drops), you're just guessing with more steps.
🙈
Ignoring Reality
The data said no — but you convinced yourself they just don't understand the vision yet. They understand. They just don't want it.
⏳
Waiting Too Long to Test
Every week you spend "getting ready" is a week you could have learned something real — and now your competitor already knows what you're about to find out.
Visual Framework
THE COMPLETE LOOP
The entire Lean Startup framework as one connected loop. Every stage, every decision, every concept. Click any node to see exactly what happens there and how to think through it.
00IDEA— something you think could work
What this stage is
An idea is just a hypothesis dressed up as excitement. The idea itself is worth nothing. What matters is what you believe is true about it — that's where the loop actually starts.
How to process an idea correctly
Wrong
"HVAC companies need an AI receptionist." → this is a solution, not a tested belief.
Right
"HVAC owners are losing bookings to missed calls and will pay to stop it." → now you have a testable assumption.
The move
Don't stay here long. The second you have an idea, translate it into the riskiest assumption and move to TEST. Ideas that sit become fantasies.
01ASSUMPTION— what you believe is true
What this stage is
Every idea contains multiple assumptions. Your job is to find the riskiest one — the belief that, if wrong, kills the entire business. Test that one first. Everything else is secondary.
How to name assumptions correctly
Too vague
"People will pay for this." → can't test it. Too broad.
Specific
"HVAC owners in the $1M–$5M range will pay $497/mo to recover missed call bookings." → testable this week.
Specific
"The loss-framing opener will convert at 2x my current demo booking rate." → measurable, binary result.
The question to ask
Write down every assumption. Then ask: "Which one, if wrong, kills everything?" That's your first test. Don't test easy ones first — test the killer.
❌ The trap
Testing safe assumptions (can I build it? can I deliver it?) while ignoring the dangerous one (will anyone pay?). Founders do this to avoid rejection. Don't.
02TEST / MVP— run the smallest possible check
What this stage is
MVP = Minimum Viable Test. Not a product. The smallest thing that puts your assumption in front of a real person with real skin in the game. If it takes more than a week, it's too much.
Real MVP types — ranked by speed
Fastest
Cold call. 30 HVAC owners. Script + opener. The call IS the test. No tech needed.
Fast
Loom demo video of SteelCity Air AI. Show it on a demo call before the system exists. Do they want it?
Medium
Landing page + "Book a demo" CTA. No product behind it. Clicks and sign-ups = demand signal.
Slower
Manual delivery. Do the service by hand for 1–2 clients. Proves willingness to pay before you automate anything.
The rule
The test only counts if a real person responds with real behavior — money, a booking, a reply, a refusal. Hypothetical responses don't count.
❌ The trap
Building the full product as your "test." You've tested nothing — you've just built. A fully working system with zero paying clients is a very expensive assumption.
03RESULT— data only. no spin.
What this stage is
Result = what actually happened. Not what you hoped. Not what you think it means yet. Just the raw behavior: numbers, replies, bookings, ghosting, objections. Write it down exactly.
How to record results — the right way
Wrong
"It went okay. A few people seemed interested." → this is a feeling, not a result.
Right
"30 calls. 7 booked demos. 12 acknowledged the pain but didn't book. 11 hung up." → this is a result.
Right
"Quoted $750 setup on 8 demos. 3 said yes immediately. 5 stalled on the setup fee." → clear signal.
What to track
Track behavior, not opinions. Did they pay? Did they book? Did they ghost? Where in the process did people drop off? That's where the real answer lives.
❌ The trap
Measuring vanity metrics — website views, email opens, social engagement. These feel like progress but confirm nothing about your core assumption.
04LEARN— translate data into meaning
What this stage is
Learn = translate the result into a verdict on your assumption. Confirmed? Partially confirmed? Dead? No spin. No "but if only." The result is the answer — your job is to read it clearly.
The three verdicts — and what each means
DEAD
0 demos from 30 calls. Assumption wrong. Change the opener OR change the problem. Run it again. If still zero after 2 rounds → kill or pivot.
PARTIAL
Pain confirmed, pricing blocked. The problem is real — your offer or price point is off. Adjust the variable and test again. Don't change the problem.
CONFIRMED
3 of 8 paid. Assumption validated. Don't stop — immediately name the next riskiest assumption and loop again.
The discipline
If your learning doesn't change your next action, you didn't learn — you rationalized. Real learning = different decision.
❌ The trap
"They just didn't understand the value." "Wrong timing." "Wrong sample." You're explaining away the data. The data is the truth. Learn from it.
05DECISION— stay, pivot, or kill
What this stage is
Every loop ends in a decision. There are only three options. You pick one, then immediately set up the next assumption and go again. Hesitation here is the same as building blind.
The three decisions — and when to make each
STAY
Assumption confirmed. Keep going. Tighten what's working. Name the next assumption. Don't over-celebrate — one confirmed assumption doesn't mean the business works.
PIVOT
Pain is real but offer/pricing/channel is wrong. Change ONE variable. Test again. Pivoting ≠ starting over. It means adjusting a specific thing and re-running the loop.
KILL
Nobody cares about the problem. You've tested it multiple ways. Still nothing. Kill the assumption entirely. Redirect energy to a different problem. This is not failure — this is efficiency.
The decision tree
Nobody cares → change the problem. They care but won't pay → fix the offer. They pay → keep going and find the next assumption to test.
❌ The trap
Pivoting too early (scared of the data) or too late (attached to the vision). Both kill businesses. The data tells you which. Read it without ego.
→SELECT A STAGE— click any node in the diagram above
Idea
Where it starts
Assumption
Riskiest belief
Test
Smallest check
Result
Raw data only
Learn
Translate data
Decision
Stay / pivot / kill
Section 08
THE CHEAT SHEET
One screen. Everything you need. Memorize this.
The Loop
ASSUMPTION → TEST → RESULT → LEARN → DECISION → REPEAT
The 5 Questions (Run These Always)
1. What do I believe is true?
2. How do I test this as fast as possible?
3. What actually happened?
4. What does that mean?
5. What do I do next?
Decision Rules
IF nobody cares →change the problem
IF they care but won't pay →fix the offer
IF they pay →keep going, find next assumption
IF unclear →run a tighter test
The One Sentence
TEST → SEE WHAT HAPPENS → DECIDE → REPEAT
Section 09
FINAL TEST
Don't skip this. Explaining it back is how you actually install it.
Question 1
Explain the loop in your own words. Don't copy it — explain what it actually means and why it matters for building a real business.
Question 2 — The Scenario
You want to start a service that does social media content for local gyms. You believe gym owners are frustrated that they don't have time to post consistently, and they'd pay $300/month for someone to handle it.
What do you do first? Walk me through your exact next move using the loop.
Question 3 — The Trap Answer
If you said: "I'd build out the service, create packages, make a website, and then start reaching out to gyms" — that's the wrong answer.
The right answer: Name the riskiest assumption. Test it this week. Don't build anything until someone says yes.
The riskiest assumption isn't "can I make content." It's "will a gym owner actually pay $300/mo for this." Test that first. Everything else is premature.