What CAC is and how to calculate it when the customer belongs to the marketplace

Every seller eventually asks the question, usually the month the ad invoice stops being small: how much does it cost me to get a customer? The question sounds simple and has a one-line arithmetic answer. The trap is that on a marketplace, that one line hides something important: the buyer who just paid you is not your customer. They are Amazon’s customer, or MercadoLibre’s, lent to you for one transaction.

CAC — customer acquisition cost — was born in businesses where the brand actually owns the buyer’s identity: their email, their history, their account. In D2C you can follow a person from the first ad to their fifth order. On Amazon, the buyer’s email reaches you masked behind an encrypted alias from the messaging system, and platform policy forbids using that thread for marketing. On MercadoLibre you receive what you need to ship and invoice, and Mercado Libre’s own privacy notice states that a seller who receives buyer data becomes the controller of that data and cannot use it for a purpose different from or incompatible with the transaction that produced it.

None of that makes CAC useless. It makes it a metric you have to redefine on purpose before you calculate it, because if you copy the D2C formula unchanged you will report a number that means nothing and then set budgets with it.

This article takes the metric apart in three pieces: what it is and what it isn’t, how to calculate it with what the channels actually give you, and what public measurements do and do not say about what a customer should cost.

what CAC is and what it is not

CAC is the total cost of getting one new customer in a period. The canonical formula: everything you spent acquiring customers during the period, divided by the number of new customers who arrived in that period.

The two words doing all the work are total and new. Total means it is not just what you paid Amazon Ads or Product Ads: it also includes the first-purchase coupon you discounted, the agency or freelancer running the campaigns, the samples you sent a creator, the photo shoot you produced specifically for that campaign. New means that orders from someone who had already bought from you do not belong in the denominator, even if they came through an ad.

That is the first source of confusion. Most sellers who say “my CAC is 180 pesos” are actually reporting cost per order or cost per conversion, which lumps first-time and repeat buyers together and counts only platform spend. That number is useful too, but it is a different metric answering a different question.

It is also not ACoS. ACoS tells you what percentage of attributed sales went to advertising; CAC tells you how many pesos it took to bring one person in. You can hold ACoS perfectly flat while your CAC doubles, simply because the same spend is now buying repeat orders instead of new buyers.

Glossary: ACoS is ad spend over attributed sales; it measures campaign efficiency, not the cost of a new customer →

the real constraint: you don’t have buyer identity

In a D2C business, “new customer” is a row in your database. On a marketplace you do not have that row, and that constraint shapes everything that follows.

Amazon acknowledges the gap and offers a partial way out: new-to-brand (NTB) metrics. Amazon defines a new-to-brand order as a purchase from someone who had not bought any product from your brand on Amazon in the previous twelve months. Divide ad spend by NTB orders and you get a cost per new brand buyer, which is the closest thing to a real CAC that exists inside the channel. It requires Brand Registry enrollment, and not every campaign type reports it: today it appears natively in Sponsored Brands and Sponsored Display, not in Sponsored Products. Amazon has been shifting availability by country and by campaign type, so confirm in Seller Central and in the Amazon Ads console what is actually enabled on your account before you build the report.

MercadoLibre publishes no equivalent metric. Product Ads reports spend, clicks and attributed sales, but does not separate first-time buyers from repeat ones. There, the approximation has to come from your own orders: if your catalog lets you recognize the same buyer across orders, you can estimate what share of the month’s sales came from someone who had bought before. That is an estimate, not an identity, and it should be reported as such.

The practical conclusion: on a marketplace your CAC will never be as precise as a direct store’s. It can be consistent, which is what you actually need to allocate budget.

what the public measurements report

This is where invented numbers thrive, so let’s be blunt. There is no public CAC benchmark for Mexican marketplace sellers. AMVO’s annual online sales study publishes market size and buyer habits — for 2025 it reported Mexican retail ecommerce at roughly 941 billion pesos, growing close to 19%, with around 77 million digital buyers — but that is market size, not a seller’s cost of acquisition.

What does exist are ecommerce measurements, almost all of them from direct-to-consumer stores, almost all of them from the United States, in dollars. Two worth reading side by side:

• Shopify published its own 2021 data for stores with fewer than four employees: roughly $21 in arts and entertainment, $127 in health and beauty, $129 in fashion and accessories, $129 in home and furniture, and $377 in electronics.
• First Page Sage, a US agency, publishes a CAC-by-industry report built from its own portfolio of 80-plus clients between 2020 and 2025: around $53 in food and beverage, $61 in beauty and personal care, $66 in fashion, $76 in consumer electronics and $91 in jewelry.

Look at electronics: $377 in one source and $76 in the other. Neither is lying. They measure different things — different samples, different years and, above all, different definitions of which costs belong in the numerator. So the honest use of these figures is as order of magnitude and as ranking between categories: jewelry and electronics cost more than food and consumables, and that ordering repeats across sources. Treating them as a target for a Mexican seller on MercadoLibre would be inventing precision that does not exist.

The other reference you will see quoted everywhere is the 3:1 LTV:CAC ratio. It comes from the SaaS and venture capital world, not from a marketplace measurement, and it is a rule of thumb about unit economics: a customer’s lifetime value should be worth at least three times what it cost to bring them in. Useful as a compass; not a benchmark for your category.

how to build your own baseline

Since the external benchmark cannot be your target, set your own. Three months of discipline is enough.

• Pick a period and freeze it. Calendar month, always the same cut-off. Mixing weeks with months ruins the series.
• Build the full numerator. Amazon ad spend plus MercadoLibre ad spend, plus first-purchase coupons and discounts, plus whoever operates the campaigns, plus the cost of content produced for them. When in doubt about a line item, ask whether you would still spend it with zero intention of bringing in new buyers.
• Build the denominator from the best data you have. On Amazon, new-to-brand orders. On MercadoLibre, orders from buyers you do not recognize from earlier months — or, if you cannot tell them apart, total orders, clearly labeled as a proxy.
• Report it per channel before consolidating. An Amazon CAC and a MercadoLibre CAC tell you where to move budget; the average of the two tells you nothing.
• Keep the series. The first month is a number without context. By the third you have a trend, and the trend is what decides.

a worked example where the math closes

Take one month on Amazon Mexico. You spent 42,000 pesos on Sponsored Brands and Sponsored Display, 6,000 on first-purchase coupons and 12,000 on the freelancer running the campaigns. Numerator: 60,000 pesos.

The campaigns reported 500 attributed orders, of which 200 were new-to-brand. Divide by 500 and you get 120 pesos, and you call it CAC. It isn’t: that is cost per attributed order. Brand CAC is 60,000 over 200, or 300 pesos per new buyer.

Is that healthy? It depends on exactly one thing: what that buyer leaves behind. If your product sells for 1,160 pesos including tax and, after referral fee, fulfillment fee, shipping and product cost, you keep 190 pesos of net margin per unit, then a customer who buys once and never returns cost you 300 and left you 190. You lose 110 pesos per new customer. If in your category that buyer comes back an average of 2.4 times, they leave 456 pesos and the math closes comfortably. The same CAC is a disaster or a bargain depending on repeat purchase, which is why CAC is never read alone.

Glossary: real net margin, with commissions, fees, shipping, tax and advertising already deducted →

how this reads in iqseller

CAC is not a metric the panel invents: it is a division whose numerator and denominator live in two different modules, and the real work is making both refer to the same period and the same SKU.

The Profitability module holds the economic side: net margin per SKU with the COGS you loaded, commissions from the Amazon settlement and from MercadoLibre orders, FBA and Full fees, shipping, the VAT and withholding breakdown, and the period’s advertising. That is the margin any acquisition cost has to be compared against, and it is the part almost nobody has at hand when deciding to raise a budget.

On the spend side, advertising from both channels lands in the same period and the same Parent → Model → SKU tree, so you can see what was invested per product family instead of an account-level total. And in Alerts, the most useful thing to watch is not CAC itself but the margin holding it up: when a SKU’s net margin drops, a CAC that was healthy yesterday stops being healthy without the campaign changing a thing.

the mistake that costs money

The expensive mistake is not miscalculating CAC. It is comparing it against the selling price instead of against the margin. A CAC of 300 pesos on a 1,160-peso ticket looks like 26% and feels reasonable; against a 190-peso margin it is a loss. The selling price is not yours — the referral fee, fulfillment, shipping, tax and product all come out of it first.

The practical rule fits in one sentence: CAC is compared against what you keep, not against what you charge, and it is judged against how often that buyer returns. Everything else — industry benchmarks, category averages, 3-to-1 rules — is context. Your number is the one that matters, and it is the only one you can genuinely build.