Cart abandonment rate: what the benchmark really is and why it barely applies on a marketplace

There is one figure that shows up in practically every piece of ecommerce content: seven out of ten carts get abandoned. It gets quoted in decks, in agency proposals and in sales emails for recovery tools. What almost never gets quoted is where it came from, which stores it was measured on, or, and this is the part that matters for anyone reading this blog, whether that metric can even be calculated on the channel you sell on.

If you sell on Amazon Mexico or on MercadoLibre, the short answer is that it cannot, because the cart is not yours. The buyer arrives through a search engine you do not control, adds to a cart you cannot see and pays in a checkout that belongs to the platform. When they abandon, nobody reports it to you. Chasing that metric on a marketplace is spending time on a number that does not exist, and sellers do try, because generic ecommerce content never draws that line.

Where it does exist, and where it is worth real effort, is in your own store: Shopify, Tiendanube, WooCommerce or whatever you use to sell direct. There the checkout is yours, every step is measurable, and abandonment stops being a borrowed statistic and becomes a concrete list of friction you can remove.

This post covers how to calculate it properly, what benchmark genuinely exists and which market it comes from, why on a marketplace you should change metrics instead of forcing relevance, and what moves the number when the channel really is yours.

what cart abandonment is, and how many versions of it exist

The standard definition is 1 minus (completed purchases divided by carts created) over a period. If 90 visitors added something to the cart in a month and 27 finished buying, your abandonment rate is 70%.

So far, simple. The problem is that in practice three different metrics with similar names coexist, and people mix them up:

• Browse abandonment: viewed a product and never added anything. It is the largest of the three and the least actionable, because a lot of those people were comparing, not buying.
• Cart abandonment: added to cart and did not buy. This is the one almost everybody quotes.
• Checkout abandonment: already started the payment process, entered details, chose shipping, and did not finish. It is the smallest and by far the most valuable, because that person had already decided to buy and something stopped them.

When someone tells you your abandonment is terrible, the first question is which of the three they are measuring. The third one points at a real problem in your store; the first one is largely how people behave on the internet.

It is worth sizing the opportunity. With 1,000 sessions, 90 carts and 27 purchases, cart abandonment is 70% and site conversion is 2.7%. If a recovery campaign brings back 5% of the 63 abandoned carts, that is about 3 extra orders a month. At a ticket of 1,391 pesos, the figure reported in 2026 by an own-store platform in Mexico and not a marketplace data point, with a 30% contribution margin, that comes to roughly 1,300 pesos a month. It is money, but it is a long way from “recover 70% of your lost sales”.

the benchmark that does exist, and where it comes from

There is a serious benchmark, and it deserves to be quoted properly. The Baymard Institute maintains a compilation of cart abandonment studies, and its published 2026 average is 70.22%, calculated as the average of around fifty separate studies. What is notable about that number is that it has stayed in a narrow band, between 69% and 71%, since 2014: it is not a figure that improves with design trends.

Two important nuances from the same material:

• The device gap is large. The collected measurements put mobile abandonment above 85% and desktop around 73%. A store with mostly mobile traffic will show a worse average than a desktop-heavy one with nothing wrong in its checkout.
• The category range is enormous. It runs from around 56% in grocery and pet to over 84% in fashion, and climbs past 90% in travel, where price comparison is normal behaviour.

Now the part that has to be said plainly: that material is almost entirely United States and European ecommerce, measured on own stores. There is no equivalent public measurement of cart abandonment for Mexican ecommerce, much less broken down by category. The annual AMVO online sales study, which is the best public reference for the national market, sizes the digital channel’s growth in Mexico, 19.2% in 2025 and 17.7% penetration over retail, but it does not publish an abandonment rate. So 70.22% works as an order of magnitude and as a reason not to panic, not as a target.

why this metric barely exists if you sell on a marketplace

This is the honest part of the post, and the one that saves the most time.

On Amazon you do not own the funnel. You do not know who added your product to a cart, how many did, or who left it sitting there for three weeks. What Seller Central gives you is sessions, page views and the percentage of sessions that ended in a unit sold. That is conversion, not abandonment. The only approximation to the cart lives in the brand reports, where the search query performance report shows how many times an item was added to cart per search term, and even there the data is aggregated by keyword and available only with brand registry. It is worth confirming which reports you actually have enabled directly in Seller Central, because Amazon changes them often.

On MercadoLibre something similar happens, and on top of that a good share of the traffic buys directly from the listing without passing through a cart at all. The metrics the platform exposes are visits, questions and sales per listing. There is no “added and did not buy” event you can see. If you also sell under models where MercadoLibre controls the full experience, the distance between you and the checkout is even greater.

The practical conclusion is not that the behaviour does not exist, of course people add and do not buy on both marketplaces, but that it is not your metric because it is not your lever. You cannot change that checkout, you cannot email that person, and you cannot measure whether what you did worked.

Glossary: real net margin, with everything deducted →

what to look at on a marketplace instead

If the cart is not yours, the equivalent measurable metric is listing conversion rate: visits that end in a sale. There you do have the number and there you do have levers, price, photos, title, reviews, availability, delivery time. What conversion rate is normal on each channel covers what to expect on Amazon and on MercadoLibre, which is not remotely the same thing.

Beyond conversion, three signals play on a marketplace the role abandonment plays in your own store:

• Unanswered questions on MercadoLibre. A question is a buyer with a specific objection and live intent. Response time is the closest thing you have to a recoverable cart.
• Conversion drops without traffic drops. If visits hold and sales fall, something changed on the page or in the competition: a price, a delivery window, a new review. That is your abandonment, and it is diagnosable.
• Availability and the buy box. A listing that is out of stock or lost the buy box converts at a fraction of normal. The buyer did not abandon your cart, they never reached it.

what actually moves abandonment when the channel is yours

In your own store the diagnosis is well documented. Baymard’s surveys of American shoppers, again not Mexican ones, produce an order that repeats year after year.

First, by a wide margin, around four in ten abandonments with a stated reason come down to extra costs: shipping, taxes and fees that appear late in the process. It is not that shipping is expensive, it is that it shows up after the person has already settled on a price in their head. Before that, a very large group, roughly four in ten respondents, was simply browsing with no intention to buy, and that group is not a problem to solve.

After extra costs, the recurring reasons are the familiar ones: being forced to create an account in order to buy, a process that is too long or has too many fields, not being able to see the total before the very end, not trusting the site with card details, delivery times that turn out longer than expected, and a returns policy that does not convince.

Translated into what you can actually do in a Mexican store:

• Show the shipping cost as early as possible, ideally before the cart, with a postal code.
• Offer guest checkout. The account can be created after payment.
• Cut fields. Every extra field is a chance to leave.
• Offer the payment methods people use here, including monthly instalments where it makes sense and cash payment options if your category calls for them.
• State the delivery window in concrete days, not vague ranges, and hit it.
• Make the returns policy visible before payment, not in a footer link.

None of that is exotic. Most recoverable abandonment falls away with transparency: the full price, on time, and fewer steps.

building your own baseline and segmenting it

A store-wide abandonment average is close to useless. The version that works has at least three cuts:

• By device. Mobile and desktop are different businesses. If your mobile abandonment is twenty points worse than desktop, your problem is the mobile checkout, not your offer.
• By traffic source. Traffic from a discovery campaign abandons far more than someone who searched for you by name. Mixing them hides both stories.
• By checkout step. This is where the value is. Knowing which screen people fall off on, shipping details, carrier choice, payment, turns a metric into a task.

And one hygiene rule: measure checkout-initiated abandonment, not just cart abandonment, and compare it against itself week over week. Comparing against a global average of 70.22% will not tell you anything you can act on Monday.

Glossary: real-time synchronization across channels →

where iqseller fits

Worth being clear about what it does and does not do. iqseller does not measure your checkout: store analytics, the step-by-step funnel and recovery campaigns live in your ecommerce platform and in your analytics tool. What the panel answers is the question that comes next, and that almost nobody answers: how much the order you rescued is actually worth.

The Profitability module calculates net margin per SKU with the COGS you loaded, the commissions from the Amazon settlement and from MercadoLibre orders, FBA and Full fees, shipping, advertising and the breakdown of VAT and withholdings. With your own channel in the same view as the marketplaces, the comparison stops being intuition: a D2C order pays no marketplace commission but does pay acquisition and shipping, and knowing which of the two leaves more pesos per order is what decides how much to invest in recovering carts.

The Inventory module keeps stock aligned between your own store and the marketplaces, which is exactly where the D2C channel most often hurts itself: a cart recovered for a product that no longer exists is a cancellation, not a sale. Catalog links the same real product across its listings and its SKU on every channel, with the Parent → Model → SKU tree, and Alerts flags when a price or an availability drifts apart between the store and the marketplace.

Put differently: you lower abandonment in your store; the panel tells you whether what you are rescuing is worth what rescuing it costs.

what to review every month

• Checkout-initiated abandonment, by device, against your own previous month.
• The exact step where most people drop, and one hypothesis at a time to fix it.
• Profit per order on your own channel against your marketplace channels, with shipping and acquisition included.
• The real cost of your recovery campaigns against the orders they brought, not against emails opened.
• Stock alignment between store and marketplaces, because every recovery that ends in a cancellation costs more than the sale that never happened.

And if you only sell on marketplaces: skip the metric and measure conversion per listing. It is the same question, asked where you actually have both the answer and the lever.