Everything I know about business

🧵 [Mega-thread] Everything I know about business:

Lessons and mental models for building businesses and companies.

Mainly a summary of Alex Hormozi's strategy. A repeatable system anyone can use to scale their business.

Warning: it's very long.

If you want the text ready to copy into your AI agent, leave me a comment 💬

The whole system can be summed up like this:

1️⃣ Choose a goal
2️⃣ Map the business
3️⃣ Find the constraint
4️⃣ Fix the constraint
5️⃣ Scale what works
6️⃣ Standardize it
7️⃣ Delegate it
8️⃣ Find the next constraint
9️⃣ …and start again

The hard part isn't learning another tactic. It's correctly identifying what matters now and ignoring everything else.

Part 1: PREPARATION ✅

1. What scaling really means

Scaling doesn't simply mean increasing revenue.

A business scales when it can increase the number of customers it serves and the profit it generates without causing a proportional increase in founder involvement (your time), costs, complexity, customer complaints, etc....

A business that doubles revenue but triples its headcount, destroys margins, increases churn, and forces the founder (you) to work twice as much.

It has grown, sure, but it hasn't scaled well.

A good north star is:

👤 Profitable customers retained per period

This combines four things that revenue alone hides:

🧲 You acquire customers.
🛒 Customers actually buy.
💎 Customers get enough value to stay.
💰 The company makes money serving them.

Growth can be boiled down to three major levers:
→ Get more customers
→ Make each customer worth more
→ Keep customers for longer

Ultimately, everything else should support one of those three outcomes or reduce the risk of them stopping.

2. Apply the framework to one project at a time

Every business has a constraint.

A portfolio of businesses also has a constraint, which will usually be one of these:

👀 your attention
💰 your available capital
👥 your ability to hire operators
🛠️ your technical ability
📣 your distribution
⚖️ your risk tolerance

Don't try to find and solve the constraint for five projects at the same time.

For each operating cycle, choose:

→ One main project that gets most of your attention
→ One proven project that stays in maintenance mode
→ At most, one small exploratory bet

Within each project, identify just one primary constraint.

You can run several projects. You can't aggressively transform all of them at once without fragmenting your efforts.

Part 2: IDENTIFY THE STAGE OF THE BUSINESS ✅

The right constraint depends on the stage the business is at.

There are 10 stages: Discover, Monetize, Acquire, Stabilize, Focus, Productize, Optimize, Organize, Specialize, and Capitalize.

The sequence is what matters:
Learn manually → prove monetization → generate demand → stabilize delivery → focus → productize → optimize → organize → specialize → allocate capital and build lasting enterprise value.

Use this practical interpretation:

1. 🔎 Discover - Is there a genuinely painful problem for a specific buyer?

❌ Typical mistake: Building a huge product.

2. 💰 Monetize - Will they pay for the outcome?

❌ Typical mistake: Branding, automation, hiring.

3. 📣 Acquire - Can we generate qualified opportunities repeatably?

❌ Typical mistake: Opening five channels at once.

4. ⚙️ Stabilize - Question: Can we serve customers consistently?

❌ Typical mistake: Buying a lot more traffic.

5. 🎯 Focus - Which avatar, offer, and channel should we focus on?

❌ Typical mistake: Adding more products.

6. 📦 Productize - Can delivery be carried out with a repeatable system?

❌ Typical mistake: Hiring without a process.

7. 📈 Optimize - Can we improve margin, conversion, retention, and capacity?

❌ Typical mistake: Constantly reinventing everything.

8. 🏗️ Organize - Can functions operate with clear owners?

❌ Typical mistake: The founder approving everything.

9. 🧑‍🔧 Specialize - Would more specialized roles outperform generalists?

❌ Typical mistake: Creating unnecessary management.

10. 💸 Capitalize - Where should the next unit of money and attention go?

❌ Typical mistake: Diversifying before mastering the core.

The stage rule:

❌ Don't solve a problem from a stage you haven't reached yet.

Part III: DEFINE THE OUTCOME ✅

Choose one business goal for the next 90 days

A constraint only exists in relation to a goal.

Bad goals:
❌ Improve marketing
❌ Improve the product
❌ Grow on social media
❌ Add AI

Better goals:
✅ Increase monthly profit from €20,000 to €35,000.
✅ Increase retained paying customers from 500 to 750.
✅ Reduce customer acquisition payback from nine months to five.
✅ Reduce monthly churn from 6% to 3%.

The template is:

Increase [business outcome]
from [current baseline]
to [target]
by [date]
without worsening [guardrails].

Example:
Increase profitable customers retained
from 100 to 150 per month
within 90 days
without reducing gross margin below 65%
or increasing first response time beyond four hours.

It's important to separate the outcome from the method!

The goal isn't: "Hire three salespeople."

Hiring is an intervention.

The real goal could be: "Increase completed sales from 40 to 70 per month."

Maybe hiring is the right move.

But maybe the real problem is:
→ Lead quality
→ Sales rep utilization
→ Follow-up
→ Pricing
→ Poor qualification

Never put the proposed solution inside the goal.

Part IV: MAP THE BUSINESS ✅

Write out the complete system from left to right

Almost any business can be represented like this:

Market
→ Offer
→ Attention
→ Lead
→ Sales conversation or checkout
→ Payment
→ Onboarding
→ Delivery
→ Customer outcome
→ Retention
→ Expansion
→ Referrals

Different business models use different labels.

💼 Services
Lead → appointment → show-up → sale → payment → service → renewal/referral

💻 SaaS
Visitor → signup → activation → paid conversion → retention → expansion

🛒 E-commerce
Visitor → product → cart → checkout → delivery → repeat purchase

🤝 Marketplace
Supply → demand → match → transaction → successful fulfillment → repeat

For each stage, measure:
📥 Input volume
🎯 Conversion rate
⏱️ Cycle time
🛡️ Safe capacity
💰 Cost
✅ Quality
👤 Owner

The basic funnel equation

End customers = Initial opportunities × conversion rate at each stage

Example:
10,000 visitors
× 5% convert into leads
× 20% book
× 70% show up
× 30% buy
= 21 customers

Improving the percentage that looks lowest isn't automatically the right move.

A stage is only the constraint if improving it increases the system's final output.

There's no point widening the mouth of a pipe if there's a bottleneck further upstream.

Part V: FIND THE CONSTRAINT ✅

Start with this test:

"If we doubled our advertising, outreach, content, or other demand-generating activity tomorrow, could we properly serve the additional customers?"

- If doubling demand creates operational chaos, you have a supply/capacity constraint.

- If the business could comfortably serve those additional customers, you have a demand constraint.

A supply constraint usually shows up in:
→ Inventory
→ Equipment
→ Qualified staff
→ Onboarding
→ Fulfillment
→ Customer support
→ Management
→ Working capital
→ Founder availability (your time)

A demand constraint usually shows up in:
→ Market selection
→ Offer strength
→ Awareness
→ Lead generation
→ Lead quality
→ Sales conversion
→ Pricing
→ Follow-up
→ Customer unit economics

The most common constraints can be grouped into four areas:

→ Getting customers
→ Selling
→ Delivering
→ How the company operates

Your time is usually underneath all four.

Now go through the full diagnostic tree:

Question 1: Is there a proven market?

→ Does a specific group experience this problem repeatedly?
→ Is it painful enough that they're already spending money or time trying to solve it?
→ Can the buyer afford a solution?
→ Can you identify and reach the buyer?
→ Do people pay, or do they just say the idea sounds interesting?

If not, the constraint is probably the market or problem selection.

If yes, keep going.

Question 2: Can you sell the offer manually?

→ Can the founder close qualified buyers?
→ Do customers understand the promised outcome?
→ Does the offer seem meaningfully different from the alternatives?
→ Does the price feel acceptable relative to the perceived value?
→ Do buyers trust that the solution will work?

If qualified prospects consistently refuse to buy, the constraint is probably in:
→ Offer
→ Positioning
→ Market
→ Sales conversation

If they buy, keep going.

Question 3: Are there enough qualified opportunities?

→ Are enough of the right people seeing the offer?
→ Are they responding?
→ Are they becoming leads?
→ Is lead quality acceptable?
→ Can volume be increased predictably?

If not, the constraint is acquisition.
If yes, keep going.

Question 4: Do qualified opportunities convert?

Measure:

🪡 Lead response
→ Qualification
→ Booking
→ Show rate
→ Close rate
→ Payment collection

The constraint could be:
→ Slow response
→ Poor qualification
→ Booking friction
→ Low show rate
→ Weak sales process
→ Lack of follow-up
→ Payment friction
→ Bad offer

If there are opportunities but no sales, the constraint is sales or the offer.

If they buy, keep going.

Question 5: Can the business deliver twice the volume?

→ At what stage would a backlog build up?
→ Where are customers currently waiting?
→ Which employees are constantly overloaded?
→ Where does quality drop as volume increases?
→ What work can only be done by one person?
→ What inventory, capital, equipment, or supplier would run out?

If something breaks, the constraint is in delivery or operational capacity.
If the system can handle more, keep going.

Question 6: Are customers getting the promised outcome?

Measure:
→ Activation
→ Time to first value
→ Completion
→ Customer outcomes
→ Complaints
→ Refunds
→ Support load
→ Repeat usage

If customers buy but don't get the outcome, the constraint is the product or delivery quality.
If they do, keep going.

Question 7: Do customers stay and increase in value?

Measure:
→ Cohort retention
→ Churn
→ Repeat purchases
→ Renewals
→ Expansion
→ Referrals
→ Gross profit per customer over time

If acquisition works but customers leave quickly, the constraint is in:
→ Retention
→ Customer success
→ Recurring value

If they stay, keep going.

Question 8: Are the economics scalable?

Measure:
→ Gross margin
→ CAC
→ Contribution margin
→ Payback period
→ Refunds
→ Working capital needs
→ Realized lifetime gross profit
→ Payment timing

If each customer ends up profitable but consumes too much cash up front, the constraint could be in:
→ Payback period
→ Working capital

If customers are structurally unprofitable, the constraint is in:
→ Business model
→ Pricing
→ Delivery cost
→ Retention

Question 9: Can the company operate without the founder?

Ask:
→ Does every decision require founder approval?
→ Do employees wait for answers?
→ Is important knowledge only in the founder's head?
→ Can someone else sell, hire, set prices, or solve customer problems?
→ Does performance collapse when the founder isn't there?

If so, the constraint is in:
→ Founder dependency
→ Management
→ Operating systems

Confirm that it's actually the constraint

A bad metric isn't necessarily a constraint.

Use four tests:

1️⃣ Counterfactual test

"If this metric improved by 50% tomorrow and everything else stayed the same, how much would the business's final outcome improve?"

If the answer is "practically nothing," it's not the current constraint.

2️⃣ Queue test

"Where is work, demand, inventory, information, or decision-making waiting?"

Persistent queues usually reveal the bottleneck.

3️⃣ Throughput test

"Which stage currently sets the lowest ceiling on the number of profitable customers completed?"

That stage is a strong candidate for the constraint.

4️⃣ What breaks next test

"If we removed this constraint, what would break next?"

A real constraint should reveal the next bottleneck once it's removed.

Checklist to confirm a constraint:

☑ Explains why the main goal isn't being reached.
☑ Improving it should increase final output.
☑ There's enough upstream input to benefit from the improvement.
☑ Downstream stages can initially absorb the improvement.
☑ The constraint can be measured.
☑ It can be tested through a specific intervention.
☑ The proposed improvement won't destroy quality, margin, or retention.

Part VI: PLAYBOOKS FOR EACH CONSTRAINT ✅

❌ Market constraint

Symptoms
→ People like the idea but don't pay.
→ Sales require very long explanations.
→ Prospects don't feel any urgency.
→ Almost every conversation ends with a price objection.
→ Retention is poor because the problem isn't important enough.
→ You can't clearly identify the buyer.
→ Acquisition channels contain few suitable prospects.

Diagnostic tests

💬 Behavioral interviews
→ Ask about what people have already done:
→ When did the problem last happen?
→ What did it cost?
→ What have you tried?
→ What are you currently paying for?
→ Who controls the budget?
→ What happens if you do nothing?

Don't rely on: "Would you buy this?"

People are bad at predicting their hypothetical buying behavior.

💰 Manual presale
Describe a clear outcome, ask for payment, and deliver the outcome manually.

Payment is much stronger evidence than:
→ Praise
→ Surveys
→ Email signups
→ Waitlists

👁️ Concierge pilot
Deliver the service manually before building the scalable product.

This reveals:
→ The real workflow
→ The required inputs
→ Common objections
→ What the customer values
→ Which parts can be automated later

🚀 Exit condition

There is repeated paid demand from a recognizable type of buyer for a recognizable outcome.

A good early-stage starting point is five to ten paying customers acquired through founder-led sales, although the level of evidence needed increases with the time and capital required to build the business. (which is what you'd expect, right?)

❌ Offer constraint

The offer isn't just the product.

It includes:
→ Customer
→ Promised outcome
→ Mechanism
→ Deliverables
→ Price
→ Payment structure
→ Proof
→ Risk reversal
→ Timing
→ Support
→ Bonuses
→ Positioning
→ Scarcity or urgency

Hormozi's course explicitly includes:
→ Market selection
→ Pricing
→ Value Equation
→ Offer construction
→ Bonuses
→ Guarantees
→ Scarcity
→ Urgency
→ Naming

🧮 The Value Equation

Hormozi's value equation is:

Value = (Desired outcome × Perceived likelihood of success) / (Time delay × Effort and sacrifice)

You need to:
🔼 Increase the desired outcome
🔼 Increase the perceived likelihood of success
🔽 Reduce the time needed to achieve the outcome
🔽 Reduce the effort required

🔎 Offer audit 🔎

⭐️ Desired outcome

Ask:
→ What does the customer actually want to achieve?
→ Are you selling the outcome or just describing features?
→ Is the outcome important enough to justify taking action?

📊 Perceived likelihood

Build confidence through:
→ Demonstrations
→ Case studies
→ Testimonials
→ Transparent methodology
→ Guarantees
→ Credentials
→ Proof of past results
→ Clear explanations of why the mechanism works

⏳ Time delay

Reduce the time until the customer:
→ Sees progress
→ Receives the first deliverable
→ Gets the first meaningful result
→ Understands that the purchase was worth it

🩸 Effort and sacrifice

Reduce:
→ Setup
→ Learning
→ Data entry
→ Coordination
→ Decision fatigue
→ Required behavior changes
→ Switching friction
→ Ongoing maintenance

🧾 The reusable template

☑ Customer:___________
☑ Problem:___________
☑ Desired outcome:___________
☑ Current alternative:___________
☑ Unique mechanism:___________
☑ Time to first value:___________
☑ Total time to outcome:___________
☑ Effort required from the customer:___________
☑ Deliverables:___________
☑ Proof:___________
☑ Risk reversal:___________
☑ Price:___________
☑ Payment terms:___________
☑ Reason to act now:___________

🧪 Offer experiment
Keep the same audience and acquisition method while testing one important offer variable:
→ Outcome
→ Price
→ Packaging
→ Guarantee
→ Payment terms
→ Time to value
→ Proof
→ Deliverable
→ Positioning

Measure both:
"Conversion rate" and "Gross profit per opportunity"

A cheaper offer may convert more prospects but generate less profit.

❌ Acquisition constraint

Hormozi's Core Four divide lead generation based on:
1️⃣ Whether the communication is one-to-one or one-to-many.
2️⃣ Whether the audience knows you or not.

The practical result is:
👤 One-to-one
People who know you → Warm outreach
People who don't know you → Cold outreach

👥 One-to-many
People who know you → Organic content
People who don't know you → Advertising

This also includes:
→ Referrals
→ Employees
→ Agencies
→ Affiliates
as ways to get other people to generate leads.

Recommended sequence for an unproven business:

🔥 Warm outreach → ❄️ Targeted cold outreach → 📲 Consistent content → 🔖 Referrals and partnerships → 💸 Paid acquisition

It's not mandatory, but it usually lets you learn much more cheaply than starting by spending big on advertising.

🧑‍⚖️ Rule of 100

The Rule of 100 is a volume discipline:

Do approximately 100 reps of the main activity per day instead of drawing conclusions from tiny samples.

Depending on the business, a primary action could be:
→ A targeted outreach message
→ A call
→ An ad variant
→ A distribution attempt
→ A piece of content
→ A follow-up
→ A sales training rep

Adapt the volume to the channel.

In high-value enterprise sales, 100 generic messages can be much worse than 15 thoroughly researched contacts.

The principle is: "Enough consistent volume, not spam."

🏥 Diagnose acquisition by stage

Impressions
→ Attention
→ Click or response
→ Lead
→ Qualified lead
→ Appointment or checkout

📉 Low reach
Likely problem: Insufficient distribution

👀 Reach but little attention
Likely problem: Weak hook or targeting

💬 Attention but no response
Likely problem: Weak message or offer

🎯 Responses but bad leads
Likely problem: Targeting or qualification

📅 Good leads but no bookings
Likely problem: Friction or follow-up

⏰ Bookings but no-shows
Likely problem: Commitment and reminders

🚀 Acquisition exit condition

You have at least one channel that can produce qualified opportunities with acceptable economics and enough consistency to plan around it.

❌ Sales constraint

Map the entire sales process

Lead received
→ First response
→ Qualification
→ Appointment
→ Show-up
→ Discovery
→ Offer presentation
→ Objection handling
→ Decision
→ Payment collection
→ Post-sale reinforcement

Measure conversion and cycle time at each stage.

Distinguish an offer problem from a salesperson problem

A useful test:
→ If the founder or best salesperson can't sell the offer to qualified prospects, suspect the market or offer.
→ If the best salesperson consistently closes but the others don't, suspect training, process, management, or hiring.
→ If prospects say yes but don't pay, suspect the payment process, urgency, decision-making authority, or follow-up.
→ If the close rate is high but profit is low, suspect price or packaging.

The CLOSER structure:
C — Clarify: clarify why the prospect is there.
L — Label: label the problem.
O — Overview: review past experiences and attempted solutions.
S — Sell: sell the destination or outcome.
E — Explain: explain and remove concerns and uncertainty.
R — Reinforce: reinforce the decision after the purchase.

Its purpose is to create a repeatable structure that can be:
→ Measured
→ Reviewed
→ Trained

🔼 Sales improvement process:

1. Record calls.
2. Review wins and losses.
3. Identify the exact stage where deals die.
4. Define observable behaviors for that stage.
5. Rewrite the script or process.
6. Role-play.
7. Test it with real qualified prospects.
8. Compare performance by salesperson, segment, source, and offer.
9. Turn the winning behaviors into a scorecard.

❌ Delivery and capacity constraint

Map delivery as a production line

Payment
→ Onboarding
→ Setup
→ Core delivery
→ Quality control
→ Customer support
→ Completion
→ Renewal

For each stage, record:

📦 Current throughput
Question: How many units per period?

🛡️ Safe capacity
Question: How many can be completed without losing quality?

⏱️ Cycle time
Question: How long does one unit take?

📚 Queue
Question: How much work is waiting?

❌ Error rate
Question: How often does work have to be redone?

👤 Owner
Question: Who's accountable for it?

🔗 Dependency
Question: What person, supplier, or tool is needed?

The first necessary stage that can't absorb more volume is usually the operational constraint.

Fix delivery in this order:

1. Eliminate
→ Unnecessary approvals
→ Redundant steps
→ Low-value customization
→ Duplicate data entry
→ Meetings that don't change decisions
→ Reports no one uses

2. Protect the bottleneck
Don't let scarce specialists spend time on:
→ Admin
→ Scheduling
→ Data cleanup
→ Routine communication
→ Work someone else can do

3. Standardize
Create:
→ Templates
→ Checklists
→ Defined inputs
→ Quality criteria
→ Decision rules
→ SOPs

4. Batch
Group similar work to reduce switching costs.

5. Automate
Only automate when:
→ The manual process works.
→ The exceptions are understood.

6. Delegate
Transfer repeatable work with:
→ Clear outputs
→ Clear quality standards

7. Hire or add infrastructure
Add:
→ Employees
→ Suppliers
→ Inventory
→ Equipment
→ Software
→ Working capital

Only after confirming that additional capacity really is the constraint.

8. Regulate demand
When capacity is scarce:
→ Raise prices
→ Limit availability
→ Prioritize valuable customers
→ Eliminate low-margin work
→ Create a waitlist
→ Turn one-to-one delivery into one-to-many where it makes sense

🚀 Delivery exit condition

The business can take on materially more customers without an unacceptable drop in:

→ Quality
→ Delivery time
→ Customer outcomes
→ Gross margin
→ Employee workload
→ Support response time
→ Refunds or complaints

❌ Retention constraint

A company can look like it's limited by acquisition when the real problem is that customers disappear too quickly.

Map the retention system
Purchase
→ Onboarding
→ Activation
→ First value
→ Habit or repeat usage
→ Outcome
→ Renewal
→ Expansion

Diagnose churn based on when it happens

🛒 Immediately after purchase
Likely problem: Incorrect expectations, buyer's remorse, poor handoff

🚪 Before activation
Likely problem: Setup friction or unclear onboarding

🧪 After initial usage
Likely problem: Weak product or poor fit

✅ After achieving the outcome
Likely problem: Product solves a finite problem with no continuity

📉 After several months
Likely problem: Declining value, competition, price, or lack of a recurring use case

🔄 During renewal
Likely problem: Poor communication of value or procurement friction

📊 Retention metrics
Use cohorts instead of overall averages.

Measure:
→ Activation rate
→ Time to first value
→ Percentage that achieves the promised outcome
→ Retention by acquisition source
→ Retention by segment
→ Gross Revenue Retention
→ Net Revenue Retention where relevant
→ Repeat purchase rate
→ Refund rate
→ Reason for churn
→ Expansion rate

🔼 Improve retention in this order:

1. Sell to customers with a better fit.
2. Set the right expectations.
3. Improve onboarding.
4. Reduce time to first value.
5. Make progress visible.
6. Make sure customers do the highest-value behavior.
7. Intervene proactively when usage drops.
8. Improve the core outcome.
9. Add continuity only when there really is ongoing value.
10. Create natural expansions once the first outcome is achieved.

❌ Economics and cash constraint

Growth is dangerous when the company doesn't understand its economics.

Main formulas:

👤 Customer Acquisition Cost
CAC = Total acquisition and sales cost ÷ New paying customers

Calculate both:
1. Paid media CAC
2. Fully loaded CAC, including:
→ Sales salaries
→ Commissions
→ Software
→ Agencies
→ Creative production

🤑 Gross profit per customer
Gross profit = Revenue − Direct cost to serve the customer

💰 Contribution profit
Contribution profit = Revenue − Direct delivery costs − Variable acquisition and sales costs

⏳ Payback period
Payback period = CAC ÷ Monthly gross profit per customer

🤑⏳ Realized lifetime gross profit
Realized lifetime gross profit = Gross profit actually collected from a customer cohort

Whenever possible, use real cohort data.

Don't build your whole plan around an optimistic LTV estimate from immature customers.

🏥 Economics diagnosis

CAC too high

Investigate:
→ Targeting
→ Conversion
→ Sales productivity
→ Channel costs
→ Offer strength
→ Price
→ Referral rate

Gross margin too low

Investigate:
→ Delivery labor
→ Custom work
→ Supplier costs
→ Refunds
→ Support load
→ Product mix
→ Price

Payback too long

Investigate:
→ Upfront payment
→ Annual plans
→ Deposits
→ Setup/onboarding fees
→ Faster activation
→ Higher price
→ Lower CAC
→ Earlier upsells
→ Lower delivery cost

Customer value too low

Investigate:
→ Retention
→ Repeat purchases
→ Expansion
→ Cross-sell
→ Continuity
→ Customer quality
→ Outcome quality

Money Model architecture

Offer architecture can be split into four types:

1️⃣ Attraction offers: make the initial purchase or interaction easier.
2️⃣ Upsells: let customers buy additional value.
3️⃣ Downsells: offer a smaller option or one structured differently.
4️⃣ Continuity offers: create recurring value and recurring payments.

A generic stack could be:

- Attraction: Low-friction entry or small sample outcome
- Core: Main solution
- Upsell: More speed, scope, convenience, access, or outcome
- Downsell: Smaller scope, payment plan, fewer features, or self-service
- Continuity: Service, maintenance, monitoring, replenishment, or ongoing access

Don't add an upsell just to extract more money. It should solve the customer's next problem or improve their current outcome.

🚀 Economics exit condition:

Incremental customers generate attractive contribution profit, the payback period can be financed, and growth doesn't cause a liquidity crisis.

❌ Operations, people, and founder constraint

Founder bottleneck test

The founder is probably the constraint when:
→ Every important decision requires their approval.
→ Employees wait instead of acting.
→ Only the founder can close important deals.
→ Customer exceptions always escalate to them.
→ Hiring stops when the founder is busy.
→ The company has no written decision rules.
→ Results get worse immediately when the founder steps away.

📦 Productize before delegating

The preferred sequence is:

The founder does the work well
→ Repeats it
→ Identifies the winning process
→ Documents it
→ Trains someone else
→ Measures their output
→ Transfers authority

Don't document every experimental process.
You'll end up with a huge library of SOPs for things that don't work.

🚂 Define roles through outputs

Bad role description:
→ Help with marketing.

Better:
→ Generate 80 qualified sales opportunities per month at a fully loaded cost of less than €100 per opportunity.

Every role needs:
→ Outcome
→ Metric
→ Authority
→ Inputs
→ Process
→ Quality standard
→ Review cadence

Delegation ladder
1. Observe: the employee watches.
2. Execute with instructions: the founder directs every step.
3. Execute with review: the employee acts and the founder checks.
4. Recommend: the employee proposes decisions.
5. Decide within limits: the employee has defined authority.
6. Own the outcome: the employee controls the process and metric.

Delegation isn't complete until decision rights are transferred along with the work.

👔 Management cadence

A manager should be accountable for:
→ A clear output
→ A small set of leading metrics
→ The people and resources needed
→ A weekly operating review
→ Corrective actions when metrics are off

Don't hold someone accountable for an outcome without giving them the authority they need to produce it.

❌ Focus and risk constraint

Once the main engine works, reduce the risks that could stop it.

Audit:
→ Founder or key-person dependency
→ Customer concentration
→ Acquisition channel concentration
→ Supplier concentration
→ Platform dependency
→ Regulatory dependency
→ Geographic concentration
→ Dependency on a single product
→ Technology or data dependency
→ Liquidity and refinancing risk

Don't diversify too early.

Diversification has a cost:
→ It fragments attention
→ It reduces volume per channel
→ It slows learning

Reduce concentration only when:
→ The main engine is already working.
→ The concentration creates a material existential risk.
→ The alternative can be developed without hurting the core.
→ There's enough management capacity to support it.

Part VII: CHOOSE THE INTERVENTION ✅

Separate the constraint from the proposed solution

Example: "Onboarding capacity is limited to 100 customers per month."

Possible interventions:
- Eliminate unnecessary onboarding steps.
- Create group onboarding.
- Automate account setup.
- Improve training.
- Hire another onboarding specialist.
- Charge for personalized onboarding.

The constraint isn't:
→ We need another employee.

That's just one possible intervention.

🧮 Intervention prioritization formula

A useful approximation:

Priority = (Expected impact × confidence × speed) / (cash required × effort × complexity)

This isn't precise math. Its purpose is to force you to explicitly compare the alternatives.

For each intervention, estimate:
→ Expected change in the final metric
→ Confidence based on evidence
→ Time to get evidence
→ Cash required
→ Founder attention required
→ Reversibility
→ Downside risk

Choose the smallest intervention capable of testing the hypothesis about the constraint.

Part VIII: RUN THE EXPERIMENT ✅

Experiment rules
→ Change one important variable at a time.
→ Use enough volume to distinguish signal from noise.
→ Run the test for at least one full business cycle.
→ Define success before seeing the results.
→ Measure both local and final metrics.
→ Include quality and margin guardrails.
→ Don't change the test halfway through because the early results feel uncomfortable.
→ Record unexpected effects.

🔎 Interpret the result correctly

1️⃣ The local metric improves and final output improves

The constraint diagnosis was probably right.

Scale the intervention.

2️⃣ The local metric improves but final output doesn't

Possible explanations:
→ It wasn't the real constraint.
→ Another stage immediately became the new constraint.
→ The local metric was a vanity metric.
→ Downstream capacity couldn't absorb the improvement.
→ Customer quality got worse.

3️⃣ Final output improves but a guardrail collapses

You've increased activity, not scalable throughput.

Examples:
→ Sales increase but refunds double.
→ More customers come in but gross margin disappears.
→ Delivery volume increases but outcomes get worse.
→ Leads increase but their quality drops.

4️⃣ Nothing changes and volume was low
The result is inconclusive.

Don't confuse a tiny sample with a failed strategy.

5️⃣ Nothing changes after sufficient volume

Stop the intervention and update your mental model.

Part IX: SCALE WHAT WORKS ✅

More, Better, New

A good practical starting point is:
70% → More
20% → Better
10% → New

7️⃣0️⃣ %: More
Do more of what already works.

Examples:
→ Increase budget on a profitable campaign.
→ Publish more versions of a proven content format.
→ Send more of the outreach that gets responses.
→ Add capacity to a productive sales process.
→ Expand a successful offer to more suitable prospects.
→ Create small variations of a winning ad.

This is usually the lowest-risk allocation.

2️⃣0️⃣ %: Better
Improve or expand what already works.

Examples:
→ Better targeting
→ Better hook
→ Better proof
→ Higher-converting landing page
→ Faster onboarding
→ Better sales script
→ Better packaging
→ More efficient delivery
→ Adjacent customer segment

It carries moderate risk because significant changes can also break what's already working.

1️⃣0️⃣ %: New
Try something fundamentally different.

Examples:
→ New acquisition channel
→ New product
→ New segment
→ New pricing model
→ New market
→ New sales motion
→ New delivery mechanism

It's reasonable to expect most new experiments to fail.

The goal isn't immediate efficiency. The goal is to discover the next winner.

🪜 Promotion ladder

NEW: A genuinely different experiment works
↓
BETTER: Improve and validate the mechanism
↓
MORE: Scale repeated, proven variations
↓
STANDARDIZE: Turn the winner into a process
↓
DELEGATE: Assign ownership

🆙 When to move from More to Better
Do it when:
→ You can't safely increase volume.
→ Marginal returns are declining.
→ The channel is reaching capacity.
→ Quality is deteriorating.
→ Additional volume is prohibitively expensive.

🆙 When to move from Better to New
Do it when:
→ The process is materially optimized.
→ Further improvements yield small returns.
→ The market or channel is saturated.
→ Concentration creates an unacceptable risk.
→ A new opportunity has a higher expected return than continued optimization.
→ Don't move to New just because you're bored.

Part X: STANDARDIZE AND DELEGATE ✅

Turn the result into an operating asset.

When an experiment works repeatedly:
→ Document the exact inputs.
→ Document the required behaviors.
→ Define the expected output.
→ Define quality standards.
→ Define common exceptions.
→ Assign an owner.
→ Train using real examples.
→ Review performance.
→ Remove the need for founder approval where possible.
→ Keep testing improvements.

A process isn't finished just because a document exists.

A process is finished when another competent person can repeatedly produce the expected result without the founder having to direct every action.

Part XI: OPERATING CADENCE ✅

1️⃣ Daily cadence

Focus on the main actions related to the current constraint.

Examples:
→ Outreach attempts
→ Sales conversations
→ Customer interviews
→ Follow-ups
→ Units delivered
→ Recruiting conversations
→ Experiment variants
→ Churn interventions

Don't confuse activity with output. But measure the activity needed to generate the output.

2️⃣ Weekly constraint review

Have a 30- to 60-minute meeting.

Agenda:
→ What's the 90-day goal?
→ What's the current result versus the goal?
→ What's the current constraint?
→ What evidence supports the diagnosis?
→ Has the constraint metric improved?
→ Has the final metric improved?
→ Have any guardrails gotten worse?
→ What have we learned?
→ What will we continue, stop, or change?
→ Who owns the next action, and by when?

Don't fill the meeting with general status updates.
Status can be written down. Meetings should produce decisions.

3️⃣ Monthly review

Review:
→ Funnel conversion
→ Cohort retention
→ Customer outcomes
→ CAC
→ Gross margin
→ Contribution profit
→ Payback period
→ Delivery capacity
→ Employee productivity
→ Founder dependencies
→ Cash needs

Ask:
→ Is the constraint still the same?

4️⃣ Quarterly review

For each project:
→ Recalculate the economic model.
→ Repeat the 2× test.
→ Identify the current stage.
→ Identify the current constraint.
→ Decide whether to scale, improve, maintain, pause, or kill.
→ Allocate founder time and capital.
→ Choose one business goal for the next cycle.

Part XII: USE THE SYSTEM ACROSS MULTIPLE PROJECTS ✅

Classify each project

🗺️ Explore
There's a plausible market, but no significant evidence of payment.

Goal: Find out whether the problem and customer are real.

🧐 Prove
Customers have paid, but the business is still delivered manually or inconsistently.

Goal: Prove repeatable demand, delivery, and customer value.

🚀 Scale
Demand, delivery, retention, and economics are sufficiently proven.

Goal: Increase volume while protecting economics and quality.

🔄 Maintain
The project works, but right now it's not the best incremental use of your attention.

Goal: Maintain performance with as little founder involvement as possible.

⏸️ Pause or Kill
The evidence doesn't justify continued investment.

Possible reasons:
→ No paid demand after the previously defined test volume.
→ Poor retention.
→ Structurally bad economics.
→ No realistic path to distribution.
→ Excessive regulatory or platform risk.
→ No available operator.
→ Another project has materially higher expected value.
→ The original thesis is no longer true.

Don't keep a project alive just because you've already put a lot of work into it.

Portfolio allocation

A reasonable adaptation of More, Better, New would be:

70%:
Proven projects and proven growth engines

20%:
Adjacent improvements and promising projects with evidence

10%:
Completely new experiments

Adapt the allocation to the stage of the portfolio.
→ A very early-stage portfolio may need more exploration.
→ A mature, cash-generating portfolio should normally allocate more resources to proven engines.

🧮 Project scorecard

Score each project from 1 to 5.

🔥 Pain
Question: How severe and frequent is the problem?

💰 Willingness to pay
Question: Is there evidence in the form of payments?

🔁 Retention
Question: Do customers keep using or buying?

📈 Economics
Question: Can each customer generate attractive contribution profit?

📣 Distribution
Question: Is there a credible way to reach customers?

🧠 Founder advantage
Question: Do you have uncommon insight, skills, an audience, or access?

⚡ Learning speed
Question: How quickly can the core hypothesis be tested?

💵 Capital efficiency
Question: How much cash is needed before getting evidence?

⚙️ Operational load
Question: How difficult is delivery?

🚀 Upside
Question: How large could the opportunity reasonably become?

Don't just add up the points.

One fatal weakness can invalidate the whole project:
→ Nobody wants to pay
→ Impossible economics
→ No legal path
→ No distribution
→ You can't deliver the product

💬 Portfolio decision

For each project, choose one:

🔂 scale: the engine works and isn't saturated yet
🔧 fix: an identifiable constraint is blocking growth
🔍 prove: important hypotheses still need evidence
🛠️ maintain: the project works but deserves little incremental attention
⏸️ pause: keep the option open while reallocating resources
🛑 kill: the thesis has failed or the opportunity cost is too high

Part XIII: UNIVERSAL DECISION RULES ✅

Use these rules when you don't know what to do next.

No clear painful problem?
→ Research problems and solve them manually.

People show interest but don't pay?
→ Fix the market, outcome, offer, proof, or price.

You can sell manually but lack opportunities?
→ Fix acquisition.

You have leads but few appointments?
→ Fix response, qualification, follow-up, or scheduling.

You have appointments but few sales?
→ Fix the offer, sales process, targeting, or trust.

You have sales but delivery is collapsing?
→ Fix capacity and operations.

Delivery works but customers aren't getting results?
→ Fix product quality and time to value.

Customers get results but leave?
→ Fix ongoing value, retention, or customer fit.

Customers stay but growth consumes too much cash?
→ Fix CAC, payback, payment timing, or the money model.

Everything works but the founder is overloaded?
→ Standardize, delegate, hire, and transfer authority.

Does the proven mechanism still have spare capacity?
→ Do More.

Does More produce diminishing marginal returns?
→ Do Better.

Are More and Better practically exhausted?
→ Try New.

Does an intervention work repeatedly?
→ Document it and delegate it.

Has the constraint disappeared?
→ Diagnose the business again.

Part XIV: COMMON MISTAKES ✅

❌ Optimizing something that isn't the constraint
Improving an unconstrained stage rarely changes final output.

Scaling before fixing retention
Pouring more customers into a leaky system creates:
→ More churn
→ More support
→ More reputational damage

❌ Buying traffic before proving the offer
Paid advertising can't permanently save an offer that qualified buyers don't want.

❌ Hiring to solve ambiguity
A new employee can't execute a process nobody understands.

❌ Automating too early
Automation makes a process run faster.
Bad processes included.

❌ Launching too many channels
Every new channel has a learning curve.

Fragmenting volume slows learning across all of them.

❌ Serving too many avatars
Different avatars require different:
→ Messages
→ Sales
→ Prices
→ Features
→ Delivery
→ Support

❌ Changing multiple variables at once
You can improve performance without learning what caused the improvement.

❌ Confusing low volume with randomness
Small samples produce very volatile results.
Do enough before drawing conclusions.

❌ Measuring revenue without gross profit
High revenue can hide:
→ Bad margins
→ Negative contribution profit
→ Reliance on theoretical lifetime value

Future retention assumptions don't pay today's bills.

❌ Discounting prices instead of creating value
Lowering prices can increase conversion while worsening:
→ Cash flow
→ Positioning
→ Delivery quality
→ Customer commitment

❌ Not defining guardrails
An experiment can improve the target metric while destroying the business somewhere else.

❌ Never diagnosing again
Solving one constraint creates the next.

The complete operating cycle:

1️⃣ Choose a project.

2️⃣ Identify its current stage.

3️⃣ Define one economic goal for the next 90 days.

4️⃣ Map the entire customer and delivery system.

5️⃣ Measure volume, conversion, capacity, quality, and economics.

6️⃣ Run the 2× demand test.

7️⃣ Identify the one stage limiting final throughput.

8️⃣ Confirm it using:
- Counterfactual test
- Queue test
- Throughput test

9️⃣ Choose the smallest intervention capable of testing the diagnosis.

🔟 Define in advance:
- Success
- Failure
- Duration
- Guardrails

1️⃣1️⃣ Run enough volume to learn.

1️⃣2️⃣ Evaluate the final business outcome, not just the local metric.

1️⃣3️⃣ Kill the interventions that fail.

1️⃣4️⃣ Scale the ones that work using More, Better, New.

1️⃣5️⃣ Standardize the repeatable winners.

1️⃣6️⃣ Delegate the process and transfer authority.

1️⃣7️⃣ Diagnose the business again.

1️⃣8️⃣ Repeat indefinitely.

The central rule is:

"Work on the one problem currently limiting the entire business. As soon as it stops being the constraint, stop optimizing it and find the next one."