In 1999, Kozmo.com raised $280 million to deliver ice cream, DVDs, and snacks in under an hour. Free delivery. No minimum order.
They had millions of customers. Explosive growth. Magazine covers.
They also lost money on every single order.
Eighteen months later, they were bankrupt.
The lesson isn't that Kozmo was a bad idea. Instacart, DoorDash, and Gopuff proved the model could work. The lesson is that Kozmo never did the math.
They confused traction with viability.
This is still the number one killer of startups today. Not bad products. Not weak teams. Founders who build businesses they don't understand.
The One Sentence That Changes Everything
A sustainable business is one where you can profitably acquire a customer, serve them at positive margin, and keep them long enough to earn back your acquisition cost with money left over.
That's it. Every other metric is a derivative of this fundamental truth.
The Four Numbers That Matter
Forget the 47 metrics on your investor deck. There are only four numbers that determine whether your business can work:
| Metric | What It Tells You | The Question It Answers |
|---|---|---|
| CAC | Cost to Acquire a Customer | How much do I spend to get one new customer? |
| LTV | Lifetime Value | How much total gross profit will that customer generate? |
| Payback Period | Months to Recover CAC | How long until I'm "even" on that customer? |
| Contribution Margin | Per-Unit Profitability | Am I making money on each transaction? |
Everything else is context. These four are the physics.
CAC: The Hidden Complexity
The formula everyone knows:
CAC = Total Sales & Marketing Spend / New Customers Acquired
The formula almost no one uses correctly:
Fully-Loaded CAC = (Marketing + Sales Salaries + Sales Tools +
Allocated Overhead) / Paid New Customers
Mistake 1: Including organic customers
If 40% of your customers come from word-of-mouth, your blended CAC looks great. But your paid CAC—the marginal cost of each incremental customer—might be catastrophic.
Always calculate two CACs:
- Blended CAC: Total spend / Total new customers
- Paid CAC: Marketing spend / Customers from paid channels only
The gap between these numbers is your organic advantage. If they're the same, you're buying 100% of your growth.
LTV: Where Optimism Goes to Die
Lifetime Value is the most manipulated metric in startup land.
The basic formula:
LTV = Average Revenue Per Customer x Gross Margin x Customer Lifespan
The honest formula:
LTV = (Monthly Revenue x Gross Margin) / Monthly Churn Rate
The second formula forces you to use actual churn, not projected retention.
The LTV Lies Founders Tell Themselves
| The Lie | The Reality |
|---|---|
| "Our customers stay forever" | You've been operating for 8 months. You don't know. |
| "Expansion revenue will kick in" | Maybe. Measure it when it happens. |
| "We'll reduce churn with product improvements" | Then reduce it first, update LTV second. |
The Magic Ratio (And When It Lies)
LTV:CAC is the ratio VCs love. Here's the conventional wisdom:
| LTV:CAC | What It Means |
|---|---|
| > 5:1 | Underinvesting in growth (or lying about LTV) |
| 3:1 | Healthy and scalable |
| 2:1 - 3:1 | Acceptable, watch carefully |
| < 2:1 | Unsustainable |
But LTV:CAC alone is dangerous. Here's why:
Scenario A:
- LTV: $3,000, CAC: $1,000
- LTV:CAC: 3:1
- Payback: 24 months
Scenario B:
- LTV: $1,500, CAC: $500
- LTV:CAC: 3:1
- Payback: 6 months
Same ratio. Completely different businesses. Company A needs 4x the working capital to grow at the same rate as Company B.
LTV:CAC tells you if your economics work. Payback period tells you how fast they work.
Payback Period: The Real Constraint
The formula:
Payback Period (months) = CAC / (Monthly Revenue x Gross Margin)
Benchmarks That Actually Matter
| Business Type | Target Payback | Why |
|---|---|---|
| SMB SaaS | < 12 months | High churn, need fast recovery |
| Mid-Market SaaS | 12-18 months | Acceptable if NRR > 100% |
| Enterprise SaaS | 18-24 months | Long sales cycles, high retention |
| E-commerce | < 6 months | Transactional, repeat purchase model |
The Payback Rule of Thumb:
Your payback period should be less than 1/3 of your average customer lifespan.
Contribution Margin: The Forgotten Metric
Revenue is not profit. Gross profit is not profit. Only contribution margin tells you if serving one more customer makes or loses money.
The formula:
Contribution Margin = Revenue - All Variable Costs
If your contribution margin is negative, you lose money on every customer regardless of scale. More growth = more losses.
This was Kozmo's problem. This was MoviePass's problem.
You cannot scale your way out of negative unit economics.
The Unit Economics Audit
Use this quarterly:
Acquisition Health
- CAC calculated by channel, not just blended
- Sales team costs included in CAC
- Trend is flat or improving
Monetization Health
- Contribution margin is positive
- Pricing based on value delivered
- Expansion revenue measured separately
Retention Health
- Churn measured by revenue, not just logo count
- Cohort analysis shows stable or improving retention
Capital Efficiency
- Payback period < 12 months (or justified)
- LTV:CAC > 3:1 (or clear path)
What to Do Monday Morning
- Calculate your true CAC — Include everything. By channel.
- Calculate 12-month cohort LTV — Use actual data, not projections.
- Calculate payback period — This tells you your cash constraint.
- Build a cohort analysis table — Monthly cohorts, track for 12+ months.
- Identify your worst channel — The one that's underwater. Consider cutting it.
The Bottom Line
Unit economics is not a finance exercise. It's a survival skill.
Every decision you make—who to target, what to charge, how much to spend on acquisition—ultimately shows up in these numbers.
Get them right, and scaling is a matter of pouring fuel on a working engine.
Get them wrong, and you're just accelerating toward a cliff.
The best founders check their unit economics the way pilots check their instruments: not because something's wrong, but because it's how you stay alive.