As of November 6, 2026, two new store formats opened within hours of each other on opposite sides of the Pacific, and both put the same operational question on the table: how much availability depth a new store can actually carry once pickup and delivery promises are attached to it. Saigon Co.op opened Co.opmart Pro 230 Nguyen Trai in Ho Chi Minh City with AI shopping trolleys, self-checkout and a Click&Collect counter. T&T Supermarket opened its first San Francisco store with 200 kinds of vegetables and a delivery and pickup plan starting in December. In both cases the shelf, not the storefront, decides whether the format works.
One format bets on the trolley, the other on the produce wall
Co.opmart Pro is positioned deliberately between the standard Co.opmart supermarket and the premium Finelife chain, which means its assortment has to look premium enough to justify the format while staying broad enough to serve a weekly basket. The store carries a food court in front of the shopping area and a Click&Collect service, so the same shelf has to serve a walk-in shopper, a Platinum member pushing a smart trolley and an online order waiting to be picked.
T&T's San Francisco store goes the other way. Its pitch is depth on fresh: roughly 200 kinds of vegetables, 17 kinds of mushrooms and more than 300 own-brand items inside 52,000 square feet, with online orders and express delivery promised from December 4 alongside in-store pickup. Depth on that scale is exactly what makes a pickup promise hard to keep, because the items that differentiate the store are also the items with the shortest shelf life and the thinnest buffer.
Why the pickup counter turns a gap into a visible failure
In a normal aisle a gap is quiet. The shopper sees an empty facing, picks a substitute and the sale is saved without anyone recording the loss. A pickup counter removes that cover: the order has already been accepted and paid, the picker walks the same aisle, and every missing line becomes a decision that has to be made in front of the customer, usually within a few minutes and often with a refund attached.
That is why stockout depth, meaning how many lines are short and how deep the shortfall runs, matters more for a new format than the headline count of SKUs. A store that lists 10,000 items but runs short on 300 of them every afternoon is less reliable than a store that lists 7,000 and keeps the promised assortment in stock. The number a format should be judged on is the cancel rate on the lines it advertises, not the size of the catalogue.
The delivery side applies the same pressure from the other direction. PYMNTS reported that Amazon's own data shows purchase rates rising more than 20% when the promised wait drops from more than four days to two, and rising another 15% or more when the wait falls from two days to hours. Walmart, meanwhile, is adding 300 drone delivery sites in the United States after more than 2 million drone deliveries. Every one of those promises is ultimately backed by a shelf that either has the item or does not.
Where new formats usually break in the first quarter
| Failure mode | How it shows up | First signal to watch |
|---|---|---|
| Depth promised on fast-turning lines | Empty facings on the produce wall or the prepared food counter by mid-afternoon | Line-level availability on the advertised assortment, measured by day part |
| Digital orders competing with walk-in demand | Orders cancelled minutes before pickup because the last unit sold at the till | Cancellations by reason, with the time of the pick |
| Autonomous front end outrunning replenishment | Self-checkout queues clear quickly while shelf gaps persist | Gap count per category per shift, not just queue length |
| Assortment breadth without depth | Many lines with one facing each, so a single sale empties a line | Days of supply per line against the replenishment cycle |
Depth is promised on the lines that turn fastest
New formats tend to advertise the categories that made them famous: fresh produce for T&T, prepared food and premium lines for Co.opmart Pro. Those are also the lines with the shortest replenishment cycles and the least tolerance for a missed delivery from the distribution centre. A shortfall on a packaged good is a lost sale; a shortfall on the produce wall is a visible empty shelf in the part of the store the format is selling.
The practical consequence is that availability has to be tracked at line level, by day part, on the assortment the format promises. Store-level totals hide the problem, because a store can report 97% availability overall while running at 80% on the twenty lines that appear in its own advertising.
Digital orders and walk-in demand compete for the same units
Once a Click&Collect counter or a delivery promise exists, the shelf serves two demand streams. Without ring-fencing, the last unit of a promoted item can be sold at the till minutes before a picker reaches the aisle, and the order is cancelled. The fix is operational rather than technical: reserve stock for accepted orders, publish a cut-off that reflects actual picking capacity, and treat substitutions as a measured outcome rather than an exception.
Both openings suggest the same test for any retailer copying the model. Pick the twenty lines the format advertises, measure them every day for a quarter, and compare cancel rates against mature stores. If the new format cannot hold those twenty lines, the trolley, the robot and the food court will not save the promise.
What to instrument before the second store opens
The list is short on purpose. Most format launches drown in dashboards before anyone can answer a simple question: on the lines we advertise, how often do we fail the customer who already paid? A single daily file, kept for a quarter, answers it faster than any new system.
FAQ
Why does a pickup counter expose stockout more than a normal aisle?
An aisle hides a gap, because the shopper simply walks past an empty facing and picks a substitute. A pickup counter does the opposite: the order was already accepted and paid, so every missing item becomes a visible failure that has to be resolved in front of the customer, usually within minutes.
What should a store network measure in the first weeks of a new format?
Line-level availability on the assortment the format promises, not the whole catalogue. Track which items were ordered, which were picked, which were substituted and which were cancelled, then compare the cancel rate against the same items in mature stores of the same banner.
Does a wider assortment always improve availability perception?
No. A wider assortment raises the number of lines that must be replenished at the same shelf capacity, so a thin replenishment cycle turns extra width into extra gaps. Depth on the items the format actually advertises matters more than the raw count of SKUs.
Sources
- Vietnam News: Saigon Co.op opens new Co.opmart Pro supermarket — store size, AI trolleys, self-checkout, sampling robots and Click&Collect.
- Supermarket News: T&T Supermarket opens first San Francisco store — 52,000 square feet, produce depth, own-brand count and the December online launch.
- PYMNTS: Amazon raises seller fees as Walmart courts holiday shoppers — delivery-promise conversion data and Walmart's drone expansion.
