Small baskets make stockouts costlier. That sentence is the whole commercial story of this festive season in Indian quick commerce, where order volumes rose 75% while gross merchandise value rose about 60% — more orders, each carrying fewer items and the same fixed cost to pick, pack and deliver. As of November 15, 2026, the operators who added dark stores fastest are the ones most exposed to the arithmetic, because every unavailable item now costs a larger share of the order it was meant to fill.
Share gains came from smaller categories
The growth did not come from big-ticket electronics. Grocery was the fastest-growing online category at roughly 40%, and more than 75% of grocery and gifting sales moved through quick commerce. Those are low-value, high-frequency items, which is exactly why the average basket fell even as the channel grew. Analysts quoted in the coverage make the point directly: smaller orders weigh on margins because each delivery costs about the same to fulfil regardless of order value.
For a brand, the implication is about availability rather than demand. When a shopper buys four items instead of nine, a single unavailable SKU removes a quarter of the basket rather than a ninth. The same fill rate that was tolerable a year ago now costs more revenue per incident, and the loss is invisible in a monthly average because it shows up as a smaller basket, not as a cancelled order.
| Metric | Festive reading | Operational meaning |
|---|---|---|
| Order volume | +75% year on year | More picks per store per day |
| Channel GMV | +60% | Value grew slower than orders |
| Dark stores | 7,153, up 34% | Coverage expanded faster than density |
| Grocery share | About 40% growth | Low-value, high-frequency mix |
Where the cost lands
It lands in three places. Picking labour rises with orders, not with value, so a store that doubles order count doubles pick time before it doubles revenue. Delivery cost per order stays flat while revenue per order falls, which compresses contribution margin. And substitutions become more expensive, because replacing an item in a four-item basket changes the customer experience far more than replacing one in a nine-item basket.
None of this argues against the channel. It argues for measuring availability at the item level and pricing the cost of a gap against the basket it sits in, rather than against a regional average. A regional fill rate of 94% can hide a top seller that is missing in half the nodes where it matters, and the average will not tell a category manager which half. Counting basket-complete orders instead of item-level fill turns the same data into a number a merchant can act on, because it only counts the orders that arrived whole. Split that figure by order-value band and the small-basket problem becomes visible without anyone having to argue about it.
Night windows are the cheapest capacity
Order timing shifted during the season, with searches concentrated in the final days before the festival and a large share of buying happening in the evening. A store staffed to walk-in patterns loses availability exactly when online demand peaks. Extending picking capacity into the evening is usually cheaper than adding a store, and it addresses the failure mode that a monthly fill rate hides.
What to instrument next quarter
Three additions to the usual dashboard are enough. First, basket-adjusted availability: the share of orders where every item was available, not the share of items available. Second, substitution rate by category, because low-value categories tolerate substitution worse than staples. Third, order-value band, so a gap in a small basket is weighted the way the customer experienced it.
Run those three for a month and the argument inside the team changes. Instead of debating whether 94% availability is good, the discussion moves to which items, in which basket sizes, at which hours, are costing the most — and that is a question operations can answer.
Assortment follows the basket, not the catalogue
A narrower core range replenished more often beats a broad range with gaps, especially where demand density is thin. The practical test is simple: if an item has not sold in a store for three consecutive weeks while its substitute sells out, the item is occupying shelf space that a faster mover could use. Cutting it is not a retreat from range; it is how the remaining range stays available.
The channel will keep growing, and store counts will keep rising. That part of the forecast is no longer contested, which is exactly why it stops being useful as a planning input. What decides who holds share inside it is narrower than any city-count figure: whether the four items a customer wanted were all there, at the hour they asked. A brand that answers that question weekly, per node, will not need a festive-season headline to tell it how the quarter went.
FAQ
Why does a smaller basket make stockouts more expensive?
Because pick, pack and delivery cost roughly the same per order regardless of value. Fewer items per order means each missing item removes a larger share of the basket.
Is declining basket size a sign of weaker demand?
No. Order volume grew faster than value, which points to more frequent, smaller purchases rather than falling demand.
Which availability metric should replace item-level fill rate?
Basket-complete orders: the share of orders where every requested item was available, ideally split by order-value band.
Why treat night hours separately?
Online demand peaks in the evening while many stores are still staffed to walk-in patterns, so availability drops exactly when it matters most.
