Revenue growth hides sins. A brand can grow top-line for years while quietly buying every incremental customer at a loss, and the P&L will not confess until the ad platforms get more expensive or the funding environment changes. Two numbers expose the truth early: CAC payback and LTV to CAC. Most DTC teams can recite both. Far fewer calculate them honestly, and fewer still run weekly decisions from them.
We run paid acquisition and storefronts as one team with one P&L, which forces these numbers to be real. Here is how we build them and use them.
Getting CAC honest
Customer acquisition cost sounds simple: marketing spend divided by new customers. The dishonesty creeps in through the definitions.
- Count all acquisition costs, not just ad spend. Agency or in-house team costs, creative production, tooling, discounts given to first-time buyers. A welcome discount is an acquisition cost wearing a promotion costume.
- Divide by new customers only. Blending returning customers into the denominator is the single most common flattery. Retargeting your own buyers and calling the result efficient acquisition is accounting fiction.
- Match the time windows. Spend this month acquires some customers next month. Over short windows this distorts CAC badly in both directions; measure over periods long enough for the lag to wash out, and be consistent.
Then segment it. Blended CAC is a vanity number. CAC by channel, by market and by first product purchased is where decisions live, because those segments differ enormously and the blend hides the losers inside the winners.
Getting LTV honest
Lifetime value suffers from the opposite disease: optimism about the future. Three rules keep it grounded.
First, build LTV on contribution margin, not revenue. A customer who generates a lot of revenue at thin margin after product costs, shipping, payment fees and returns is worth far less than the revenue suggests. LTV must be the cash the customer actually leaves behind.
Second, use realised cohort data over a defined horizon rather than a projected lifetime. Twelve or twenty-four month contribution per cohort is measurable and falsifiable. A "lifetime" projection stretching five years into the future is a spreadsheet believing its own extrapolation. If your brand is young, say so in the model: your early cohorts are the only truth you have.
Third, watch the cohort curves, not the average. The shape of repeat purchase over the months after first order tells you whether retention is improving or decaying across cohorts. An average LTV can stay flat while your newest cohorts quietly deteriorate, and by the time the average moves, you have two quarters of weak customers on the books.
Payback: the number that governs scaling
LTV to CAC tells you whether the machine is profitable eventually. CAC payback tells you whether you can afford to run it. It is the number of months until a cohort's cumulative contribution margin covers its acquisition cost, and it governs how fast you can scale, because until payback, growth consumes cash.
Two brands with an identical LTV to CAC ratio can be in completely different positions: one recovers its acquisition cost within the first order, the other waits a year for repeat purchases to catch up. The first can reinvest immediately and compound; the second needs working capital to grow and is fragile to any rise in acquisition costs or dip in retention.
This is why payback, not the ratio, should set your scaling speed. Decide the longest payback you can finance comfortably, and treat it as a hard budget constraint per channel and market. Channels inside the constraint earn more budget; channels outside it must fix their economics or shrink, however good their platform-reported metrics look.
The weekly operating rhythm
Unit economics only matter if they change decisions on a schedule. The rhythm we run:
- Weekly: new customer CAC by channel and market against target, first-order contribution margin, early cohort signals from the most recent cohorts.
- Monthly: cohort curves updated, payback recalculated per channel, budget reallocation decided from the numbers rather than from channel managers' enthusiasm.
- Quarterly: the LTV model re-based on realised data, pricing and shipping policy reviewed against contribution margin, and the payback constraint itself revisited against the cash position.
The discipline sounds bureaucratic and takes about an hour a week once the data pipeline exists. The alternative is discovering in the annual accounts what the weekly numbers would have told you in March.
The mistakes that flatter the numbers
We audit growth operations regularly, and the same distortions appear again and again: platform-reported conversions treated as truth, so every channel claims the same orders and blended CAC looks better than reality; returns and payment fees missing from margin, inflating LTV; first-time discounts booked as marketing-neutral; and paused brand search treated as saved acquisition cost when much of it was harvesting demand that existed anyway. Each distortion is small. Together they can turn a loss-making acquisition engine into a dashboard full of green.
Key takeaways
- CAC includes all acquisition costs and only new customers; segment it or it is decoration.
- LTV is realised contribution margin per cohort over a defined horizon, not a projected lifetime.
- CAC payback, not the LTV:CAC ratio, should set your scaling speed and channel budgets.
- Put the numbers on a weekly and monthly rhythm. Unit economics that do not change decisions are just reporting.
Growth is a financing problem wearing a marketing costume. The brands that internalise that scale calmly through conditions that force everyone else to retreat.