Dark store throughput: the number brands should plan against
Platforms have stopped adding dark stores as fast as they were and started squeezing more orders out of the ones they have. That shift changes what they want from your catalogue.
- Orders per store per day is the metric platforms are now optimising. Your assortment is judged against it.
- Shelf space in a dark store is genuinely scarce, so slow SKUs get delisted faster than on a marketplace.
- Sales per facing, not total sales, is what keeps you listed. Plan packs around that.
- Expect harder assortment conversations and fewer easy listings as platforms sweat existing capacity.
For several years the quick commerce story was store count. Platforms competed on how many dark stores they had opened and how many cities they covered, and brands planned around a network that kept getting bigger.
That phase is maturing. Expansion continues, but the emphasis has moved towards getting more out of the footprint already built: more orders per store per day, higher basket values, better picking efficiency, denser delivery routes. Fixed costs spread across more orders is a more reliable path to profitability than opening more fixed costs.
This is a platform strategy shift, but it lands on brands, and most brands have not adjusted their planning to it.
Why throughput is the metric that matters
A dark store is a fixed box with a fixed number of picking locations and a fixed staffing model. Its economics are dominated by whether it processes enough orders per day to cover that fixed base.
Every decision the platform makes flows from that. Which categories deserve space, which SKUs hold their facings, how deep the assortment goes in a given catchment, what gets promoted. A product that generates plenty of gross sales but occupies four facings to do it is worse for the store than a product generating slightly less from one.
Throughput per store also varies widely between platforms, which is why the same assortment strategy does not travel. A network running high order volumes per store at a lower average basket is optimising differently from one running fewer, larger baskets, and what each wants from your range differs accordingly.
What changes for your assortment
When the network was expanding quickly, getting listed was comparatively easy because new stores meant new space. Growth from existing stores means the shelf is full. Adding your SKU now requires removing something, and category managers become far less willing to experiment.
Three consequences follow.
Delisting gets faster. Underperforming SKUs used to linger. Now they are removed, because the space has a measurable opportunity cost.
Breadth becomes a liability. A brand with twelve SKUs of which four move well and eight move slowly looks worse than a brand with four fast SKUs, even at similar total revenue, because the average per facing is what gets reviewed.
New launches get harder. Asking for space for an unproven product is asking the category manager to accept a known loss for an unknown gain. You will increasingly need to earn launches by demonstrating velocity elsewhere first.
Plan around sales per facing
The practical reframe is to stop reporting your quick commerce business in total revenue and start reporting it in sales per facing per store per day. That is close to how the platform sees you, and it produces different decisions.
Under that lens, rationalising your range is usually the highest return action available. Cutting the slowest third of your SKUs typically raises your average, strengthens your position in the next assortment review, and frees your own operational attention. It also reduces the number of listings that can go out of stock and drag your availability score.
It feels like shrinking. It is usually the opposite: a smaller, faster range holds more space over time than a broad, uneven one.
Pack architecture is a throughput decision
Pack size is often treated as a pricing question. On quick commerce it is an operations question.
A pack that suits a ten minute occasion moves quickly, gets replenished predictably and holds its facing. A large pack designed for a monthly grocery shop moves slowly in this channel regardless of how well it performs in modern trade, because it does not match the mission the customer is on.
The same logic applies to bundles. A multipack that lifts basket value without occupying additional facings is attractive to the store, because it raises value per pick. That is a genuine argument you can make in a commercial conversation, and it is more persuasive than asking for space on brand grounds.
Availability is now a competitive weapon
When shelf space is contested, an out of stock is not merely a lost sale. It is evidence, sitting in the platform’s own data, that your product does not deserve the slot.
Brands consistently underestimate how much this compounds. A fortnight of poor availability in a set of high volume catchments depresses your velocity numbers exactly when the platform is deciding the next assortment cycle, and you then argue for space from a weakened position with data that contradicts you.
Holding availability in your best catchments is worth more than chasing presence in marginal ones. Concentration beats spread when the metric is per store performance.
What to do this quarter
Rank every SKU by units per store per day rather than total sales. Identify the bottom third and decide deliberately whether each earns its place or should be cut before the platform cuts it for you. Ask your category contact where your key lines sit against the category median, and treat a vague answer as a warning.
Then concentrate replenishment discipline on the catchments where your velocity is strongest, and stop spreading thin inventory across a wide footprint to look bigger. In a network being optimised for throughput, looking bigger is worth considerably less than performing better in the stores that count.