Repeat purchase rate: what benchmark exists and how little control you have on a marketplace

After two or three years of selling, most sellers start to suspect something their reports never confirm: that a share of their sales comes from people who bought before. They sense it from the messages, from the repeat reviewers, from the customer asking “when is the small size coming back?”. But the moment they try to attach a number to it, they hit a wall. The Amazon panel does not tell them how many customers came back. Neither does MercadoLibre’s. And the order spreadsheet has names and addresses that sometimes match and sometimes do not.

Repeat purchase rate —what percentage of your customers buys from you more than once— is one of the metrics that weighs most on the health of a product business, and also one of the hardest to measure from inside a marketplace. That tension is what this article is about: what the number means, what gets published as a reference, where those references come from, and what you can do when the channel hides half the information from you.

The cost of not measuring it is concrete. Without repurchase, all your growth has to come from new customers, and new customers are bought with advertising. A business that retains can ease off the advertising pedal without sales falling; one that does not retain is tied to its ad budget forever. The difference between those two businesses does not show up in this month’s sales report. It shows up in the year’s margin.

There is an extra trap: many sellers assume their repurchase rate is low simply because the marketplace never shows it to them. Working from that assumption, they stop investing in product quality and post-sale experience, and the prophecy fulfills itself.

what it is exactly and how it is calculated

The standard definition is simple: customers who bought two or more times in a period, divided by the total number of customers who bought in that period, expressed as a percentage. If 1,000 people bought from you over the half year and 240 of them placed two orders or more, your repeat purchase rate is 24 percent.

It looks trivial until you pick the period. And that choice carries almost the entire interpretation problem. Measured over a month, a low rate does not mean people are not coming back: it means your product’s natural cycle is longer than thirty days. A school backpack is not repurchased monthly; a consumable is. Measuring repurchase with the wrong window is the fastest way to reach a false conclusion.

The practical rule is to measure over a period that contains at least two expected purchase cycles for your category. If your product is consumed in 60 days, measure over 180. If it lasts three years, repeat purchase rate is simply not your headline metric, and that is fine: plenty of businesses run on customers who buy once and recommend.

There is a second decision: what counts as a customer. In your own store it is the email or the account. On a marketplace you have neither, and that is where the uncomfortable part begins.

what numbers ecommerce publishes and which market they come from

Compilations published through 2025 and 2026 by retention-tool vendors and ecommerce consultancies agree on an order of magnitude: the average ecommerce repeat purchase rate sits around 28 percent, with most stores landing between 25 and 30 percent. Those same sources flag anything below 20 percent as a warning sign and anything above 30 percent as evidence that a retention engine is actually running.

It is essential to say where that comes from. These are figures from owned stores, mostly on Shopify, and mainly from the US market. They are not from marketplaces, they are not from Mexico, and they are published by companies that sell retention software — which does not invalidate them but does demand some distance. There is no public, auditable repurchase benchmark for sellers on Amazon Mexico or MercadoLibre Mexico.

The data point worth more than the average is the spread by category. The same compilations report ranges running from roughly 10 percent in luxury goods to over 60 percent in grocery and food, with subscriptions higher still. With that much spread, the 28 percent global average is useless to anyone as a target. It is useful for understanding that repurchase is defined above all by what you sell, not by how well you sell it.

Glossary: sales velocity, the real pace at which each SKU moves →

the underlying problem: the customer is not yours

Here is the structural difference between selling in your own store and selling on a marketplace, and no tool can erase it.

When someone buys on Amazon, Amazon considers that person its customer. It does not hand you their email address, and its policies forbid using the messaging system to pull the buyer off the platform. MercadoLibre operates on the same logic: the relationship with the buyer is administered by the channel. You receive an order, a recipient name and a shipping address, and you have to work with that.

Amazon has opened in-platform tools for brand-registered sellers to reach followers and repeat buyers with messages Amazon sends on your behalf, but contact details never change hands. Both channels revise these policies often, so whatever applies to your account is worth confirming in Seller Central and in MercadoLibre’s help center before building any plan on top of it.

The consequence is hard to accept: you can have an excellent repeat purchase rate and never find out. Your product may be the one people come back for every season, and the channel will not report it to you. That blindness is not a flaw in your operation; it is the marketplace business model working exactly as designed.

what you can measure with what the channel does give you

Not being able to measure repurchase perfectly is not the same as not being able to measure anything. There are three reasonable approximations, each with its limit stated up front.

• Matching on shipping address. It is imperfect —people move, there are offices, there are gifts— but the same address with the same name across two orders months apart is a strong signal. Treat it as a lower bound: what you detect this way is a floor, not a ceiling.
• Repurchase at the product level, not the customer level. If a consumable SKU holds a flat sales curve for months without your ad investment rising, that is retention expressing itself in aggregate even though you cannot name it person by person.
• What you control outside the channel. An insert in the package inviting warranty registration, an owned store where the customer does leave you an email, a direct support channel. None of it replaces the marketplace’s data, but it builds a base that is genuinely yours, which is the central argument in from marketplace to D2C.

An important caution about the first point: cross-referencing buyer data demands care. You are handling third parties’ personal data, which carries legal obligations in Mexico; before building any process that consolidates customer information, run it past whoever handles your legal matters.

why the repurchase you cannot see is still charging you

Even when you cannot measure it precisely, repurchase is affecting your numbers every day, in directions worth recognizing.

If your real rate is high, your effective acquisition cost is lower than you think, because part of your sales required no new advertising. You are more profitable than your arithmetic shows, and probably more conservative than necessary with your ad budget.

If your real rate is low, the opposite happens and it is worse: every month starts from zero. Your growth equals exactly your ability to buy traffic, and the day cost per click rises in your category, your sales fall in proportion. A business like that can look healthy for years and break in a quarter.

There is a third effect that almost never gets named: returns are the direct enemy of repurchase. A customer who returns rarely comes back. That is why return rate is not just a cost of the month, it is a leak of future value; we treat it in depth in the hidden cost of returns.

how to set your own baseline

Since the imported benchmark does not apply and the channel will not compute it for you, the serious path is the usual one: build an internal reference and compare against yourself.

Start by defining the right window for your category and do not move it. Then pick an approximation —address matching is the most accessible— and apply it identically each quarter. What you are after is not a correct absolute number, it is a series: if your detectable match rate went from 11 to 15 percent over six months using the same method, something is improving even if the true value is another.

Then anchor that series to the decisions you can actually make: product quality, listing accuracy, delivery time, claim resolution. Those are the levers that move repurchase on a channel where you cannot run email marketing. On a marketplace, retention is built with operations, not campaigns.

Glossary: real net margin, with everything deducted →

how this reads in iqseller

It is worth being precise about what iqseller does and does not do. The panel does not identify customers and does not compute your repeat purchase rate: that data does not come out of the marketplace, and no tool can invent it. What it does is surface the aggregate signals where retention shows up, and place them next to the money.

The Profitability module shows net margin per SKU using the COGS you loaded, the commissions from the Amazon settlement and from MercadoLibre orders, FBA and Full fees, shipping, and the VAT and withholding breakdown. Alongside that, the Inventory and Forecast modules expose sales velocity and its trend per SKU. A product whose velocity holds month after month while ad spend falls is the fingerprint of repurchase in aggregate, and it is a reading no native panel assembles for you because none of them holds both series together.

The Parent → Model → SKU tree matters more here than it looks. Repurchase is rarely for the exact SKU: the customer who left happy comes back for another size, another color, or another model in the same line. Looking at the model level rather than the loose SKU, that pattern appears. Looking SKU by SKU, it disappears into the noise.

Think of a seller running a backpack line on both channels. Grouping by model, they see their main reference holding sales velocity across the half year while ad investment falls by half, and their net margin per unit rising accordingly. They do not know those customers’ names and never will. But they know that model sells itself, and that was the decision that mattered: restock it with confidence and move the ad budget to the line that actually needs it.

Repeat purchase rate is a powerful metric and, on a marketplace, structurally incomplete. The honest move is neither to pretend you have it nor to give up because you cannot have all of it: measure what is measurable, declare the limit of the method, and use the aggregate signals that margin and velocity do give you. And if retention is going to be the axis of your business, the conversation stops being about metrics and becomes one about channel strategy, which is exactly where combining marketplace and D2C stops sounding like theory.