As of November 6, 2026, the clearest signal from MercadoLibre's third-quarter report is not the revenue line but the square meters. Net revenue and financial income rose 45% to $10.75 billion in the quarter ended September 30, while income from operations fell to $676 million and the operating margin dropped to 6.3% from 9.8% a year earlier. The company added fulfillment capacity of almost 20% quarter over quarter, roughly double the 10% it added a year earlier, ahead of the Brazilian peak season. For teams that monitor on-demand and same-day delivery economics, the useful question is what that capacity buys: more availability per catchment, or a heavier shipping bill per order.
The capacity number that sits behind the margin
Three numbers sit on top of each other. Capacity grew almost 20% sequentially, GMV reached $23 billion with 38% growth, and items sold rose 36%. Volume growth therefore ran ahead of space growth but not far enough to hold the margin, because the cost of pushing parcels through a newly opened network arrives before the density that makes it cheap. Management framed the quarter as a deliberate trade: more free shipping and more space now, better unit economics later. That framing is verifiable, and the real question is whether availability per node improves without a matching rise in cost per shipment.
Free shipping is a catchment decision
The company lowered its free-shipping threshold in Brazil, and the effect shows up in buyer behaviour rather than in basket value. Items per buyer rose 12% year over year across the region; in Brazil the pace went from single digits before the lower threshold to 16% in the quarter. That is the mechanism worth measuring: a lower threshold does not make each basket bigger, it makes the same buyer return more often. For a store network, frequency is the metric that pays for space, because a node serving repeat small orders amortises its fixed cost faster than one serving occasional large ones.
Where the sixty basis points came from
Shipping-related factors accounted for about 60 basis points of sequential margin pressure. The list is specific: a lower free-shipping threshold, fulfilment expansion, higher energy prices and temporary recruitment costs. Marketing added another 40 basis points, taking sales and marketing to 11.6% of net revenue, driven by Brazil and by an affiliate programme whose orders grew about 240% year over year and whose network passed 12 million affiliates. These are the line items that separate a capacity story from a discount story: the company is not buying volume with markdowns, it is buying frequency with delivery promises.
Three markets, three different capacity problems
| Market | FX-neutral GMV growth | Items sold | What the gap suggests |
|---|---|---|---|
| Brazil | 36% | 45% | Frequency-led: items grow faster than value, so delivery cost per real of GMV matters more than basket size |
| Mexico | 21% | 25% | Slower consumption, steady buyer growth; capacity is being added ahead of demand rather than behind it |
| Argentina | 40% | 19% | Value-led: fewer, larger baskets, so free-shipping thresholds bite differently than in Brazil |
Reading the three columns together changes the planning question. In Brazil, where items grow faster than GMV, each additional parcel carries less revenue, so the network has to be efficient at small orders. In Argentina, where GMV grows twice as fast as items, the average basket is expanding and the constraint is weight and volume rather than drop count. A single regional capacity plan cannot serve both. The Brazilian number is also the one that has now run above 40% items growth for five consecutive quarters, which means the density assumption behind the new space has been holding.
The unit economics check to run before the peak
Management said local cost per shipment in Brazil fell 5% year over year in local currency. That is the number to track against capacity, because it separates a network that is scaling from one that is simply spending. The sequence below is the practical version of that check.
The mistake to avoid is treating square meters as a proxy for capability. Space only converts into availability when it sits inside a catchment with enough order density to justify daily routes, and only converts into margin when the cost per drop falls as volume rises. That is why the interesting figure in this quarter is not the 20% but the 5%: one measures ambition, the other measures whether the ambition is starting to pay.
Three things to watch next quarter
The fourth quarter will settle whether the capacity was early or late, and the answer will show up in service levels long before it shows up in guidance. Three indicators carry most of the information, and all three are observable from outside the company. None of them requires a subscription to the earnings call.
- Availability during peak weekIf promised dates hold while cost per shipment stays flat, the new space is working; if dates slip, the constraint was density, not capacity.
- Items per buyer after the peakFrequency bought with free shipping has to survive the removal of peak promotions, otherwise the margin cost was a one-off purchase of volume.
- Marketing share of revenueAffiliate orders growing 240% is cheap until commissions compound; watch whether the 11.6% share falls back once peak ends.
For brands selling through the marketplace, the practical implication is about stock placement rather than advertising. Nodes that gained space will tilt assortment toward fast-moving, small-parcel items, and sellers whose SKUs sit in that band will see better delivery promises and, in turn, better conversion. Sellers of bulky goods should expect the opposite: the same network will route their inventory to fewer, larger hubs, and the delivery promise will lengthen.
FAQ
Does a 6.3% operating margin mean the logistics bet failed?
No. It means the cost arrived before the density. The real question is whether cost per shipment keeps falling while availability holds through the peak.
Why does items growth matter more than GMV growth here?
Because delivery cost is incurred per parcel. When items grow faster than GMV, the network is handling more, smaller drops, which is a different cost problem from larger baskets.
Is a lower free-shipping threshold a discount?
It is closer to a frequency investment. It changes how often the same buyer returns rather than how much a single basket is worth.
What should a seller do with this?
Check where the new capacity sits, then align assortment and stock depth to the nodes that can now promise same-day or next-day delivery.
