Profit After CAC

Most arguments about whether advertising is working can be settled by one subtraction: contribution margin per order, minus cost to acquire that customer.

Both numbers come from your own data. Neither requires an attribution model to interpret. And unlike ROAS, the result is denominated in money, so it can be compared directly against what the business needs.

The first-order test

Take contribution margin per order — after product cost, shipping, payment fees and packaging. Subtract acquisition cost per new customer, calculated honestly.

If the result is positive, each new customer arrives having already paid for themselves, with something left toward fixed costs. Growth is self-funding, and scaling is a question of how much volume the channel can absorb before costs rise.

If it is negative, every new customer is financed from somewhere else — existing customers, savings, or a credit line. That can be a deliberate strategy. It cannot be an accident, because the loss scales precisely with success.

The example above sits between the two: $7.08 of profit on a first order carrying $45.08 of margin. Sixteen percent survives. That is a store with very little room, where a two-point margin decline or a five-dollar rise in CAC flips it negative.

Why the first order is not the whole answer

For any store with repeat purchase behaviour, judging on the first order alone understates what a customer is worth.

Repeat orders carry no acquisition cost. The margin on them is often better than the first, because welcome discounts do not apply, though sometimes worse where loyalty offers do. Either way, they arrive without the largest cost line the first order had to absorb.

In the example, 0.67 repeat orders per customer over twelve months adds $31.62 — more than four times the first-order profit. The customer is worth $38.70 rather than $7.08.

That is a genuine and important difference. It is also the point where most stores start reasoning about customers they hope to have rather than customers they measured.

Using the retention argument honestly

Three things have to be true before extending your acquisition ceiling on the strength of repeat purchases.

The repeat rate is measured by cohort. Customers grouped by first-order month, tracked forward. Not a store-wide average, which is dominated by older cohorts that have had years to buy again and will always look better than a recent one.

The repeat margin is measured too. Second orders can carry loyalty discounts, smaller baskets, or free shipping that first orders paid for.

The payback period is funded. If a customer takes seven months to repay acquisition cost, seven months of spending has to be financed from somewhere while it happens. Growing faster makes the gap larger, not smaller — which is the specific mechanism by which fast-growing stores run out of cash while profitable on paper.

With those three, bidding above first-order break-even is a financing decision with a known shape. Without them, it is a story, and the tell is that nobody can say what repeat rate the plan depends on or when it would become obvious that it was wrong.

Payback period, and why it often matters more

Lifetime value gets the attention; payback period governs whether you can act on it.

A store funding growth from its own cash flow is constrained by how long its money is tied up, not by the eventual return. Two customers each worth $120 over two years are very different propositions if one repays acquisition cost in six weeks and the other in nine months.

The practical target is a payback period shorter than your cash conversion cycle. When acquisition cost is recovered before the associated inventory has to be repurchased, growth is self-funding. When it is not, every additional customer widens the working capital gap, and the business needs external funding to grow — regardless of how healthy the twelve-month figure looks.

The skew nobody adjusts for

Averages assume a distribution that customer value does not have.

In most stores a small proportion of customers account for a large share of repeat revenue, while the majority never return at all. A mean twelve-month value of $38.70 can be produced by every customer contributing roughly that much, or by eighty percent contributing nothing and twenty percent contributing five times as much. Those are entirely different businesses, and the average describes both.

The distinction matters because acquisition strategy should follow it. Where value is heavily skewed, the question stops being “what is a customer worth” and becomes “can we acquire more of the ones that repeat” — a targeting and product question rather than a bidding one. Paying up to the average across all channels overpays for the segments that never come back.

The check is straightforward: for one cohort, plot cumulative contribution margin by customer decile. If the top decile carries more than half the total, you are running a skewed book and the average is doing you no favours.

Watching it over time

Both inputs move, usually in the same direction.

Contribution margin per order falls when discounting deepens, when free-shipping thresholds are crossed more often, or when the mix shifts toward cheaper products. Acquisition cost rises as spend scales into less responsive audiences.

Plotted together, the two lines converge. The moment they cross is the moment additional growth starts destroying value, and it is visible weeks before it reaches net profit — which is the entire reason to track them as a pair rather than reading the bottom line once a month.