As of 6 November 2026, Assaí's third-quarter release reads as two stories at once(PR Newswire). Recurring net income reached R$248 million, up 55.2% year on year, while last-mile sales rose 204% across 104 stores(Grocery CPG). Read together they describe one stock of goods serving a shopper in the aisle, an app order reserved from the backroom and a courier arriving within the hour, and the gap between those three flows is where availability actually gets decided.
What the release actually measures
The quarter is best read as a margin story with a distribution problem attached. Gross revenue reached R$21.3 billion, up 2.7%, while gross margin widened 0.45 percentage points to 17.2% and operating expenses grew 4.2%, below the 4.4% inflation print(Foodservice News). Traffic rose 2.1% to more than 40 million customers a month, yet same-store sales moved only 0.1%, which is what trade-down looks like when shoppers keep visiting but change what they put in the basket. Nine store openings accounted for 3.6 percentage points of expense growth against 0.6 points from comparable operations, so the cost line is still being driven by expansion rather than by the existing estate(Securities.io).
Where the 204% lands
Last-mile growth of that size does not spread evenly. It concentrates in the 104 stores that carry the operation, which means roughly one store in seven is now running two picking rhythms on one sales floor. A picker walking the aisle for an app order competes with the shopper standing in front of the same facing, and the loser is usually the shopper who arrives ten minutes later. That is not a logistics failure so much as an availability accounting failure: the shelf says three units left, the picker takes two, and the third customer is told the item is out of stock even though no one stole anything.
The shelf is the shared inventory
Assaí Farma closed the quarter with 11 units and a target of 25 by December, and its e-commerce runs inside the Meu Assaí app with in-store pickup as the default option(Grocery CPG). Supplements were pushed into 300 stores, and private label moved from 140 to a year-end target of 250 SKUs. Every one of those moves adds facing-level complexity to a store that is also picking orders, which is why distribution coverage and on-shelf availability start to diverge in exactly the stores that look best on a channel dashboard.
Availability is a store-level number, not a chain-level one
A chain average hides the stores where the pickup promise breaks. The table below keeps only figures the release states, so it can be defended in a review without re-deriving anything. Every row can be traced to a published line, which is the only reason a weekly operations review should carry a table at all. Anything that cannot be traced belongs in a working file, not in the review pack.
| Metric | 3Q26 | What it changes for availability |
|---|---|---|
| Gross revenue | R$21.3bn, up 2.7% | Volume is flat, so availability gains cannot come from growth |
| Gross margin | 17.2%, up 0.45pp | Mix is being managed deliberately, not by clearing stock |
| Same-store sales | Up 0.1% | Basket substitution is already happening inside the store |
| Monthly traffic | Above 40m, up 2.1% | More visits per shelf foot, same stock |
| Stores running last mile | 104 | The measured sample for pickup availability |
| Pharmacies | 11, target 25 | New categories that need their own facing discipline |
Two of those rows matter more than the rest. Flat same-store sales with rising traffic means the average basket is thinning, and a thinner basket is more sensitive to a single missing item, because the shopper has fewer substitutes in the trolley to absorb the gap. The 104-store pickup footprint then multiplies that sensitivity, since a pickup order is rejected outright when one line is missing rather than partially fulfilled.
Six checks before the next node opens
What follows is the sequence we would run before adding a store to the pickup network, ordered so that each step can be stopped without wasting the next one. The order is deliberate: the first two steps produce data, the middle two turn data into a decision, and the last two keep the decision honest. Skip the first two and the rest becomes opinion.
- Fix the unit of measure: count availability at facing level per store per day, not at chain level per week.
- Separate demand: tag pickup reservations apart from walk-in sales so the same unit is never counted twice.
- Measure the gap: compare shelf count with system stock after each picking wave, and log the delta rather than the average.
- Score substitution: record which replacement the customer accepted, because accepted substitutions are recoverable demand.
- Set a promise ceiling: cap the pickup SKU list at what the store can actually hold, not at the chain catalogue.
- Review weekly: read the availability loss figure beside the pickup cancellation rate, since one explains the other.
The order matters. Teams that start with the promise ceiling and work backwards end up shrinking the catalogue to fit a number nobody has measured, while teams that start with the unit of measure usually find that the ceiling was never the constraint. The weekly review is the step most often skipped, and it is the only one that turns availability into a number a store manager can act on before the next wave.
Why October matters more than the quarter
October same-store sales rose 3.9%, the best reading in 15 months, tied to the chain's 52nd anniversary campaign(Securities.io). A campaign-driven lift is also an availability stress test, because promotional depth pulls traffic toward a narrower set of items than the assortment was planned for. If the pickup network keeps expanding while that lift fades, the availability question stops being about capacity and becomes about whether the promotional calendar and the pickup SKU list are planned by the same team.
FAQ
Does last-mile growth of 204% mean the channel is now material?
Not on the evidence published. The release gives a growth rate across 104 stores but no absolute revenue, so the base cannot be sized from public information alone.
Why treat availability as a store-level metric?
Because pickup promises are made and broken store by store. A chain average can look healthy while a handful of stores carry most of the cancellations.
What is the cheapest first step?
Tagging pickup reservations apart from walk-in sales. Without that split, every downstream availability figure double-counts the same unit.
Sources
- PR Newswire, Assaí reports 55% profit surge and deleverages in Q3 (2026-11-06)
- Grocery CPG, Assaí posts 55% profit surge, expands private label to 140 SKUs (2026-11-07)
- Foodservice News, Assaí posts 55% profit jump, cuts leverage to 2.15x in Q3 (2026-11-07)
- Securities.io, Assaí Q3 recurring profit climbs 55% as leverage falls to 2.15x (2026-11-07)
