The Unit Economics Audit: 3 Numbers That Tell You If Scaling Kills You
A founder told me his LTV to CAC ratio looked amazing. He wanted to double the budget and scale.
I asked one question. How long does it take to get the acquisition cost back?
He paused. Then he said, “But we earn it back eventually, right?”
Eventually is a hope with a spreadsheet attached.
That gap is where scaling kills companies. Startup Genome studied more than 3,200 startups and found that 74% of high-growth internet startups fail from premature scaling. Not from bad products. From volume applied to a model that could not carry it.
The ratio that founders trust is a lagging blended average. It reports what already happened, across everyone, on revenue he has not collected yet.
Three other numbers decide whether scaling compounds or accelerates the damage: CAC payback period, contribution margin per cohort, and the shape of the LTV curve. LTV to CAC is a scoreboard. These three are instruments. Run them before the budget increase, not after.
1. CAC payback period, measured against cash
LTV to CAC asks whether a customer is worth more than they cost. Payback asks when. The second question is the one that kills you.
The median B2B SaaS payback is 16 months. The top quartile does it in six months or fewer. The bottom quartile takes 24 months or more.
Treat that as context, not a target. Your target is set by your cash.
The test: is your payback period shorter than the runway funding it?
A 20-month payback on 30 months of runway is a business. The same payback on 12 months of runway is a financing decision in a growth costume. You are not scaling. You are betting the next round lands before the cash runs out.
💡One practical rule. Measure payback on cash collected, not on booked revenue. Annual prepay and monthly billing can produce an identical ratio and completely different survival odds.
2. Contribution margin per cohort, not blended
Blue Apron gave investors two numbers in its Q3 2018 filing.
The top 30% of customers produced more than 80% of first-year net revenue. Their contribution paid back acquisition cost in under six months, at 2.5x.
The other 70% returned 0.2x.
Blended, that portfolio reads as acceptable. Split, it shows that most of the marketing budget bought customers who never broke even in a year. Scaling does not fix that mix. Scaling buys more of the 70%.
Two rules keep this number honest.
- Use net profit, not revenue. Andreessen Horowitz names the error directly: “A common mistake is to estimate the LTV as a present value of revenue or even gross margin of the customer instead of calculating it as net profit of the customer.”
- Include the cost to serve. Support tickets, payment fees, returns, delivery, onboarding time. Contribution margin is revenue minus every variable cost that customer creates.
Then cut the cohort by channel and by signup month. If one segment carries the average, you do not have healthy unit economics. You have a cross-subsidy.
3. The shape of the LTV curve, not its value
Most LTV numbers are extrapolations. You take a few months of retention, fit a decay curve, project it forward, and call the total LTV.
Sequoia’s framing is the useful one. A retention curve either flattens, which means a real group found lasting value, or it continues to decline until almost no users are left.
Those two shapes can produce the same LTV in month three and opposite outcomes in month eighteen.
Until one cohort has flattened, you do not have an LTV. You have a trend line.
Blue Apron’s own filing shows the drift. It states that customer payback in the year after acquisition “has extended in recent cohorts as increases in the Net Contribution per Customer have been more than offset by increases in the Cost per Customer.” The curve moved. The headline ratio did not.
I built an LTV prediction model for an e-commerce platform to solve exactly this. Predicting the curve per customer cohort, instead of extrapolating one average, improved ROAS by 75% while we scaled growth 1.5x. Same spend logic. Better instrument.
4. Run the three in order, because each failure means something different
The checks are sequential. Where you fail tells you what to fix.
- Payback fails. Your problem is cash timing, not product. Fix payment terms, deposits, or prepay incentives before you touch the budget.
- Contribution margin fails. Your problem is mix or cost to serve. Cut the losing segment, reprice it, or reduce what it costs to deliver.
- Curve shape fails. Your problem is retention, and no acquisition budget solves retention. Stop scaling.
Honestbee is the Southeast Asian version of ignoring all three. TechCrunch reported internal figures showing the Singapore company lost nearly $6.5 million in December 2018, on around $2.5 million in net revenue for the month, against nearly $12.5 million of gross merchandise value. Volume was never the missing ingredient. Every extra order widened the hole.
💡Key Takeaway: Scaling multiplies the model you already have. If contribution margin per cohort is negative, growth is just a faster route to the same ending.
Final Thoughts: growth only compounds when the model underneath it already works
LTV to CAC reports the game after it is played, averaged across everyone who played. It is the last number to move and the easiest to flatter.
The three numbers here work differently. They tell you what happens at the next increment of spend. That is the only question that matters before a budget increase.
Run the audit before your next scale decision:
- Is CAC payback shorter than the runway funding it?
- Does the bottom 70% of each cohort cross contribution breakeven inside year one?
- Has any cohort’s retention curve flattened yet?
Three yeses mean scaling compounds. One no means scaling accelerates whatever is already broken. The founder who wanted to double his budget had two nos and did not know it.
If you are about to raise spend and cannot answer all three, that is the work. Book a discovery call or connect with me on LinkedIn, and tell me which of the three you cannot answer yet.
A note before you close this tab. The fact that you read this far tells me something. You already sense that the way you’ve been thinking about growth might be incomplete. That instinct is worth following.
Mervyn Chua is a growth-transformation consultant helping founders and CEOs build the strategic clarity and systems to grow in an AI-first world. If this raises questions worth exploring for your brand, let’s talk.
