What a Customer Should Cost You

A customer should cost you no more than the contribution margin on their first order, plus whatever share of a second order you can actually count on inside the period you are willing to wait for the money. That is the ceiling. It is arithmetic you can do in twenty minutes from your own numbers, and it will not match anybody’s published benchmark, because it is not supposed to.

I read the pages ranking for this term before writing. All of them publish a table of average acquisition costs by industry, all of them repeat the three‑to‑one lifetime value ratio, and then all of them pivot to a list of tactics for spending less. Not one derives a number you can hold a media buyer to. A benchmark tells you what other people pay. It cannot tell you what you can pay, because it knows nothing about your margin, your shipping cost or your bank balance.

If you want the version of this conversation that ends in a plan rather than a formula, that is what I do as an ecommerce consultant. What follows is the part you can do without me.

What should a customer cost you?

Less than your first‑order contribution margin, if the business has to fund itself from each order. Somewhere above it, up to the margin you expect from repeat orders inside your payback window, if you have the cash to wait.

Those are two different businesses, and the choice between them is a financing decision rather than a marketing one. A store paying for ads out of last month’s revenue has a hard ceiling at the first order. A store with a credit line or a raise behind it can pay above that ceiling deliberately, because it is buying a customer base and has the runway to hold the position. The failure mode is doing the second thing while believing you are doing the first.

Why is the benchmark the wrong number to start from?

Because a published average is built from companies whose margins you cannot see, and margin is the entire input. Two stores with identical acquisition costs can be a healthy business and a dying one, and the difference is in the gross margin line neither benchmark reports.

There is a second problem, quieter and worse. Benchmarks are usually built from whoever was willing to report a number, which skews towards companies comfortable with theirs. And they are almost always a blended figure across every channel, so a store buying most of its customers from paid social is comparing itself against companies that get half their orders from email and returning buyers.

The ratio the same pages recommend has the same flaw. Three dollars of lifetime value for every dollar of acquisition cost is a reasonable rule of thumb for a business that already knows its lifetime value with confidence. Most stores under a few years old do not. They have a cohort a year deep and a projection, and the projection is doing the work the evidence should be doing.

How do you calculate your own CAC ceiling?

Start with contribution margin per order, not gross margin, then decide how much of a second order you are willing to count. Here is the whole calculation on a store with an eighty dollar average order and sixty percent gross margin.

Every figure in that table is arithmetic on assumptions I have labelled, not a claim about your store. Replace all ten lines with your own and the structure survives. The point is the shape: the honest ceiling on an eighty dollar order at sixty percent margin is around forty percent of revenue, and it is roughly two thirds of the gross margin figure a store would have used if it stopped one line early.

Note what this does to the target return on ad spend people quote. A $32.55 ceiling on an $80 order is a break‑even ROAS of about 2.5, and the $40.69 ceiling is about 2.0. If someone has been running to a target of 3 without doing this arithmetic, they have been leaving volume on the table. If they have been running to 1.5, they have been losing money on every order and calling it growth.

The costs that get left out, and what they do to the ceiling

Discounts, returns handling and the second shipment. Each one is small enough to ignore individually and large enough collectively to move the ceiling by a fifth.

• Discounts and codes. If a quarter of orders use a ten percent code, that is two and a half percent off the top line and rather more off the contribution. Welcome popups and abandoned cart flows both deal in codes, so this is rarely zero.
• The cost of a return, not the lost margin. A return costs the outbound shipping, the return label, the inspection and often the resale value. Treating a return as simply an order that did not happen understates it every time.
• Exchanges and reships. Cheaper than a return and not free. Whatever your rate is, it comes out of the same line.
• Apps and platform fees that scale per order. Subscription apps priced per order or per shipment belong in contribution, not in overhead.
• Gift wrapping, inserts and samples. Anything that goes in the box.

What does not belong in the ceiling: rent, salaries, agency retainers, software that costs the same at any volume. Those are fixed costs that contribution margin has to cover in aggregate. Loading them into a per‑order calculation produces a ceiling so low that no acquisition looks viable, which is its own kind of wrong answer.

Blended CAC or paid CAC — which one are you managing?

Blended CAC is total marketing spend divided by total new customers, including the ones who found you anyway. Paid CAC is what a specific channel costs to produce a customer. You manage the second and you survive or fail on the first.

The gap between the two is the honest measure of how much demand exists without advertising. A store whose blended number is far below its paid number is being carried by organic, email and word of mouth, and the paid channel is buying a smaller share of the business than the platform dashboards suggest.

If you only change one thing after reading this, make it the third row. An account optimising towards all purchases will happily spend to reacquire people who were coming back regardless, and the reported return on ad spend looks fine the entire time. This is the same structural problem I describe in what Google Ads actually costs a Shopify store, and it costs more than any bid adjustment will ever recover.

How long can you wait to get the money back?

As long as your cash position allows, and not one day longer. Payback period is the constraint that decides whether the higher ceiling is available to you at all, and it is the variable none of the ranking pages mentions.

The mechanism is simple and unforgiving. Spend today, get the first order today, wait for the second. Every day between those events is working capital you have to fund. A store paying up to a ninety day payback is carrying roughly three months of acquisition spend on its balance sheet at all times, and that number grows exactly as fast as the store does. Growth consumes cash, and this is the line it consumes it through.

What to do when the ceiling is below what you are paying

Raise the ceiling before you cut the spend. The ceiling is made of margin, and margin has more levers on it than the ad account does.

1. Check the arithmetic on a recent ninety days, not a year. Shipping rates, card fees and cost of goods all move. A ceiling built on last year’s costs is usually generous.
2. Move average order value before you touch bids. Every dollar added to the order at the same margin rate goes almost entirely into the ceiling. Thresholds, bundles and the post‑purchase offer are all cheaper than a percentage point of acquisition cost.
3. Attack the deductions. Shipping and returns are the two largest lines below gross margin for most stores, and both are negotiable in ways bids are not.
4. Separate new from returning in the ad account. Until that exists, you do not know whether the acquisition cost is real.
5. Then, and only then, cut the channels above the ceiling. By this point you know which ones they are and by how much, which turns a panicked pause into a decision.

Cutting first feels responsible and frequently is not. Spend is the only lever that reduces revenue the same day you pull it, and a store that cuts to hit a ceiling it calculated wrongly has bought itself a smaller business for no reason.

Who should not hire anyone for this

If the arithmetic above shows your ceiling is under about fifteen dollars, outside help is not your problem and will not be your solution. At that ceiling the constraint is the product economics, and no amount of account management fixes a margin that thin.

The same goes if you have not yet done the calculation. Bringing in a consultant to optimise towards a target nobody has derived is paying someone to hit a number that may be wrong, and the engagement will produce a confident report either way. Do the twenty minutes first. It is the cheapest work in this entire post and it changes what every subsequent decision is measured against.

Where somebody like me earns the fee is after the ceiling exists and the channels disagree with it — when the paid number and the blended number are telling different stories, or when the account has never distinguished a new customer from a returning one. That is a structural problem, and it is the kind of thing most of what I write about ecommerce strategy keeps returning to.

Questions people ask