As of November 9, 2026, Swiggy's September-quarter results show Instamart added 69 net dark stores to reach 1,240 outlets across 133 cities, while Eternal's Blinkit added 308 net stores in the same three months. Revenue from operations rose 36.1% year on year to 7,571 crore rupees, yet Instamart's adjusted core loss widened to 895 crore from 849 crore a year earlier and its contribution margin slipped to negative 1% of gross order value from negative 0.2% in the previous quarterUrban Acres on Swiggy results. For teams that own distribution coverage, the reading is blunt: a wider catchment is not the same thing as more sellable stock, and the margin gap shows up on the shelf long before it shows up in the warehouse.
What 69 net stores actually buy
A dark store is a promise with a lease attached. Each new catchment converts an assortment into a public commitment: if an item is listed, a customer inside the polygon can order it and expect it within minutes. That is why network growth and inventory depth have to be planned together. Sixty-nine stores across 133 cities does not mean sixty-nine equal bets. The marginal store in a dense metro competes for the same rider pool and the same demand as the store two kilometres away, while a store in a new city carries the full cost of creating the habit.
Contribution margin is an availability metric
The number that moved the wrong way is the one that matters most. Contribution margin measures everything that happens before corporate costs: picking, packing, rider time, discounts and, critically, the orders that die at the pick. A short pick is not a lost line. It is a cancelled basket plus a support ticket plus a customer who learns not to rely on the promise. When order value grows 34.8% and margin still deteriorates, the likely answer is that the assortment grew faster than the replenishment behind it.
| Failure mode | How it shows up | First signal to watch |
|---|---|---|
| Depth promised on fast-moving lines | Empty facings by mid-afternoon | Line-level availability by day part |
| Digital orders competing with walk-in demand | Cancellations minutes before pick | Cancellations by reason and hour |
| Breadth without depth | Single-facing lines empty after one sale | Days of supply against replenishment cycle |
Where the loss actually accumulates
Three cost pools absorb the difference, and they behave differently. Discount cost is discretionary: you can stop it next week. Rider cost scales with distance and density: it improves as catchments overlap, then worsens as they cannibalise each other. Stockout cost is the quiet one, because it is booked as a support ticket rather than a marketing line, and it compounds into habit loss. Retailers that treat on-shelf availability as a weekly report rather than an hourly control tend to see the third pool grow while the first two look under control.
The promise list is an operational document
A promise list is not a marketing asset, it is a working constraint that tells the picking team which lines may never be substituted silently. Written properly, it names the lines that carry the format, sets a minimum depth for each, and defines what happens when depth breaks: hide the line from the app rather than cancel at the pick. That single rule converts a visible failure into a smaller basket, which is cheaper than a cancelled order and far cheaper than a customer who stops trusting the ten-minute promise.
Blinkit's 308 and the density trap
Adding more than four times as many stores as the nearest competitor in one quarter is a statement about capital, not about operations. Density helps until it does not: overlapping polygons raise the chance that two stores hold the same slow-moving item at half the turns, while the fast-moving lines still run short at peak. The practical countermeasure is to give every store a promise list rather than a catalogue. A store that lists 7,000 items and keeps them available beats a store that lists 10,000 and runs short on 300 every afternoon, because the customer only experiences the lines that were promised and then missed.
The ordering matters as much as the metric. Availability failures on fast-moving lines are a planning problem, competition between digital and walk-in demand is a staffing problem, and breadth without depth is a catalogue problem. Fixing them in that order costs less than fixing them in the order complaints arrive, because the first one determines whether the other two are even measurable.
What to measure before the next quarter
| Metric | Why it matters | Cadence |
|---|---|---|
| Line-level availability | Drives cancellations and contribution margin | Hourly by store |
| Days of supply on promised lines | Catches breadth without depth before customers do | Daily |
| Cancellation rate by reason | Separates stockout from rider delay | Daily |
| Store payback by cohort | Tests whether new catchments are real demand | Monthly |
None of these require new infrastructure. They require a decision about which lines are allowed into the promise, taken before the quarter starts rather than after the loss is reported. In a market where the two largest players added 377 net stores in three months, the winner will not be the one that opened the most doors, but the one whose doors still have the promised stock at seven in the evening.
FAQ
Is adding dark stores still worth it if margins are negative?
Only if the added catchment carries demand that existing stores cannot serve. Overlapping polygons raise fixed cost without raising availability, which is the fastest way to widen a negative contribution margin.
Why does availability matter more in quick commerce than in e-commerce?
A two-day promise gives a store time to recover from a short pick. A ten-minute promise does not. The same stockout that costs an online order a delay costs a quick-commerce order the entire basket.
What is the first metric to fix?
Line-level availability on promised lines, measured hourly. It is the only metric that maps directly to cancellations, and cancellations are the cost that customers remember.
