Fifteen minutes is not a speed setting. It is a radius, and a radius decides which products can even be promised. Noon's 11.11 sale runs from November 5 to 13 with discounts of up to 90 per cent, and the markdown is the headline; the operationally interesting number is the delivery promise attached to it, with noon Minutes at 15 minutes and prices from Dh1, Supermall at 90 minutes, and Amazon answering through Amazon Now and two-hour Rush. What the two retailers are competing on is how much of the catalogue sits close enough to a customer to be offered that fast, which is a question about store footprint rather than discount depth as of November 11, 2026.
Two promises, two different networks
The two promises are not the same product. A 15-minute window is a guarantee about distance; a 90-minute window is a guarantee about a pick-and-pack queue. The first is fixed when the site is chosen, the second when the shift is planned. Retailers that report both under one fast-delivery line end up measuring the wrong thing at the end of the week, and the error compounds across a nine-day sale.
Amazon's version shows the same structure from the other side. Khaleej Times reported that fresh produce, meat, dairy and frozen food carry discounts up to 50 per cent with delivery in minutes through Amazon Now, while Prime members get free delivery on Amazon Now orders above Dh25 and free two-hour Rush above Dh100. The threshold is doing the work: below Dh25 the promise is minutes, above Dh100 the promise is two hours, and the gap between them is exactly the gap between a convenience trip and a weekly shop.
A 15-minute basket is structurally small
Fifteen minutes covers roughly one to three kilometres in a dense city, and that radius narrows the assortment before any buyer touches it. Essentials, chilled grocery, beauty minis, a charger, a gift bag: items a shopper needs once and cannot wait for. A 15-minute promise on a large appliance is a demonstration rather than a habit, because holding that unit in a micro node costs more than the order returns.
This is where store selection turns into a data problem instead of a real estate problem. A candidate node's trade area has to be scored on order density, basket mix and the pickup cost of the SKUs that would actually be promised at 15 minutes. A high-potential store for a weekly shop is not automatically a high-potential node for a fast promise, and the two lists overlap less than operators assume when they reuse one site model.
Ninety minutes is where the basket grows
The 90-minute window behaves differently because it can carry a real basket: household goods, small electronics, a home item that will not fit a scooter box. The binding constraint moves from distance to picking capacity, and picking capacity is a staffing decision that only becomes visible in the last three days of a campaign.
In practice the two promises compete for the same pickers. When 15-minute orders spike they jump the queue because the clock is shorter, and the 90-minute promise starts to slip first. That is why the failure mode of a fast-delivery sale is rarely the headline promise; it is the second-tier promise that quietly misses while the dashboard still looks green.
At peak, the constraint is the picker
Discount depth is visible and gets reported. Picking capacity is invisible and decides whether the promise holds. A node that assembles 40 orders an hour in a normal week does not assemble 90 in the final weekend without a second shift, and a second shift is a decision taken two weeks earlier, long before anyone knows how deep the markdowns will go.
- Node capacity — orders assembled per hour per node, measured in the 19:00 to 22:00 window rather than as a daily average.
- Promise integrity — share of orders delivered inside the promised window, split by 15 and 90 minutes, because a blended on-time rate hides which promise is breaking.
- Substitution rate — lines replaced or dropped at picking, the earliest signal that shelf availability inside the node has broken before the customer sees it.
- Basket per promise — average value of a 15-minute order against a 90-minute order; if the two converge, the fast window is eating the bigger basket.
These four numbers also explain why a static site list ages badly. A node that looked strong on catchment population can be weak on order density, and a node that looked marginal can be carrying the highest-value baskets because of what sits around it. The list has to be re-scored after a campaign, not before it.
Measure the promise, not the markdown
A sale like this one is usually reviewed on discount depth and gross merchandise value. Both are recorded by the platform and neither says whether the network held. The fields worth keeping are the ones that survive the week and can be compared with the next window.
| Field | What it captures |
|---|---|
| Node ID | the micro site that assembled the order |
| Promise | 15 or 90 minutes |
| Assembled at | pick completion timestamp |
| Delivered at | handover timestamp |
| Substitution | line replaced or dropped at pick |
| Shelf gap | SKU unavailable at pick, flagged before handover |
Once these sit in one table, the footprint question becomes answerable: which nodes miss the 15-minute promise repeatedly, and whether that is a radius problem or a staffing problem. The first is fixed by moving or adding a node, the second by a roster. Confusing the two is the most common way a quick commerce promise dies quietly, one late order at a time. A stockout that surfaces at the pick hour, an hour before the promise expires, is still cheap to fix; the same stockout found at handover is already a failed promise in a customer's hands.
FAQ
Does a 15-minute promise mean the whole catalogue is available that fast?
No. It means the part of the catalogue held in the node serving that radius. Everything else is offered on the longer window.
Why does the 90-minute promise fail before the 15-minute one?
Because the shorter clock jumps the picking queue. When volume spikes, the longer promise absorbs the delay first.
What is the most useful metric during a sale week?
Promise integrity split by window. A blended on-time rate hides which promise is breaking and in which node.
How should store selection change for quick commerce?
Score the trade area on order density and the pickup cost of the SKUs that would be promised fast, not on catchment population alone.
