The Supermarket That No Customer Is Allowed to Enter

Retail Property Doing a Different Job

a dark tent is a retail location frequently a former supermarket or purpose-built equivalent that is closed to customers and operates solely to fulfill online orders for delivery or collection

Inside it looks like a store with aisles and shelves stocked in familiar designs because pickers move through it as shoppers would. The difference is that the only people inside work there

The concept became prominent during the rise of online grocery ordering when retailers discovered that fulfilling online orders from commercial stores degraded customers' in-store experience and led to poor pickup productivity

Both halves of that sentence are costs and are worth separating. The customer experience cost is a staff member with a large trolley blocking an aisle at rush hour. The productivity cost is that the same staff member cannot follow a straight route because the store is designed to sell to searchers rather than be harvested efficiently. A supermarket is optimized to slow you down. A warehouse is optimized to speed you up. Asking a building to do both jobs at once means it doesn'tneither of them does well

Three Ways to Fulfil an Online Grocery Order

The trade is visible in the table. Automation offers the best picking economics and is located far from customers which increases delivery costs and lengthens delivery times. In-store picking is located closer to customers and produces the worst picking economics. The dark store is the middle position

Why the Middle Position Can Win

Delivery is the dominant cost in online supermarkets and is driven by drop density that is how many deliveries a van can make per hour. Density depends on how close the deliveries are to each other and to the point of origin

A dark store located within a dense urban area serves a small radius which means short driving distances many deliveries per route and short lead times. A distant automated center can select an order for a fraction of the labor cost and then spend much more to carry it across a city

The dark store also turns a property with declining retail productivity into an asset with a different use which matters when a retailer has a store that underperforms the retailer but is located exactly where its online customers live

That last point is the one that is underestimated. The scarce input for urban compliance is not automation or labor. It is a right-sized building in the right location with a rent set years ago. A grocer with an older store already has that inventory of locations. A pure competitor has to rent it at current prices

Drop Density, With Illustrative Numbers

The numbers make the trade-off happen faster than the description. Nothing that follows is a real operator. Treat each figure as illustrative and round

Imagine a van on a four-hour overnight route. Loading and configuration consume part of that window. What's left is divided between driving between deliveries and the time spent at each door with only the second of those producing a delivered order

In a dense urban radius the drops are located a few streets apart. The driving time between them is small so the vast majority of the route is dedicated to making deliveries and the van completes a large number of orders. Spread the same number of customers over a wide suburban region served by a distant center and the driving time between deliveries increases the time at the door remains the same and the order count per route decreases

Now put a cost on it. The van the driver and the fuel are practically fixed along the route. If that fixed cost is spread over a large number of orders the delivery cost per order is modest. If the same cost is spread over half as many orders the delivery cost per order roughly doubles

Let's compare this to the picking side where an automated center actually picks for a fraction of the labor cost of a human walking the aisles. The question for any given market is simply which effect is greater. In a densely populated city savings on delivery usually win which is why the dark store persists in exactly those places and loses to automation in regions where customers are dispersed

The Rapid Delivery Experiment

The model was taken to the extreme by companies promising ten or fifteen minute grocery delivery from small dark urban stores sometimes called micro fulfillment centers or dark convenience stores

The economics of that version proved very difficult. Extremely short promise times require dense coverage of small sites each with rent staff and inventory and each serving a very small area. Average order values ​​in convenience shopping are low the delivery cost per order is high and the density required to cover fixed costs is often not achieved

The sector consolidated markedly after a period of heavy investment. What survived tended to be operations with higher value orders longer delivery times or integration into a larger retail business that could absorb fixed costs

Why Ten Minutes Was the Wrong Promise

The failure is worth dismantling because it was not a failure of execution. It was arithmetic that could not be executed and put into practice

A ten-minute promise sets the service radius before anything else is decided. A van or passenger can only cover so much ground in that time so the catchment area is fixed and small. Everything else stems from that limitation rather than any business choice

A small basin is home to a small number of households. Those households place their orders at their own pace which is not controlled by the operator. Therefore the maximum orders per site per day is limited by geography and the site has to cover its rent its staff and its inventory outside of that limited volume

The margin per order is the other half of the squeeze. Convenience baskets are small because the proposition is to buy what you forgot rather than the weekly shop. A low value basket carries a low absolute gross margin and still needs a full delivery trip

Therefore unit economics had to be rescued by the volume that geography had already limited. Adding sites did not solve the problem either since each new site divides the same basin and carries its own fixed costs

Compare that to a dark store that serves a wider radius in a scheduled window. The window is the release valve. Giving up immediacy allows the operator to group orders into efficient routes which is precisely the density that the ten-minute model had defined

The Inventory Nobody Mentions

Pickup and delivery get all the attention in this debate and there is a third cost that moves along the same axis and quietly resolves many arguments

Each storage point has its own inventory. Two sites covering a city have two sets of the same range and each set must be deep enough to avoid running out of anything a customer might order. That duplication is working capital stored in buildings and it grows roughly with the number of sites rather than the number of customers

Groceries make this clearer than most categories because much of the variety is perishable. A can that goes unsold for a week is a financing cost. A tray of fresh produce that hasn't sold for a week is a loss

Waste behaves in a specific way that matters here. It increases with the number of places where it has stock and decreases as the throughput at each place increases. A site that renews its fresh range several times a day sells the product before it expires. A site that serves a small catchment area with a low order volume maintains the same range in the face of many fewer sales and a significant portion of it goes to waste

That's the ten-minute problem again coming from a different direction. Dense coverage of small sites means many storage points each underperforming and each with its own fresh variety. The model multiplied exactly what it needed to focus on

It also explains a design choice that looks like cost-cutting and isn't. Rapid delivery operations were run in deliberately narrow ranges a few thousand lines rather than tens of thousands in a supermarket because a narrow range is the only way to keep throughput per line high enough to get fresh stock moving before it spoils

So decentralization costs money for two reasons not one. It gives up picking efficiency and multiplies inventory and waste. Delivery savings have to outweigh both

The Property Question

For a landlord converting commercial space for compliance raises real issues that are often not resolved in existing leases

Use clauses Retail leases typically require the premises to operate as a store open to the public and a dark store may violate these. Rental billing Provisions that calculate rent based on in-store sales become meaningless when the store does not record in-store sales even though it may be handling enormous volume. And in shopping centers a unit that does not generate footfall hurts neighboring tenants who pay for the center's ability to attract visitors

Planning and zoning Classifications differ between retail and distribution use in many jurisdictions and conversion may require consent that is not automatic particularly when increased van movements affect residential neighbors

These are the reasons why a conversion that makes obvious operational sense often takes longer than expected

When the Lease Was Written for a Shop

The turnover rental problem deserves its own look because it is a clear example of a contract term that fails to measure what it was written to measure

Revolving leasing exists so that the landlord shares in the tenant's success. It works when the store is the point of sale as the sales made in the building are a fair indicator of the value the location offers. Turn that store into compliance and the building can be more productive than ever and record essentially zero turnover. The revenue is real it is simply recorded in an online channel rather than at the address

The footfall problem in a shopping center is similar. The tenants of a center pay through rent and service charges for the flow of visitors that the center collectively generates. A unit that occupies the facade and does not bring in visitors is taking advantage of that shared asset without adding anything which is a direct transfer from the other tenants

None of these make the conversion wrong. They mean that the conversion is a renegotiation not a change of operations and the party that has to agree is the owner whose economics just changed

What It Means for Store Economics

A converted store changes the metrics used to judge it. Sales per square foot the standard retail measure becomes irrelevant. The relevant figures are orders fulfilled per site per day picking units per labor hour and delivery cost per order

A dark store that performs well on those measures can generate revenue comparable to a commercial store in the same building with a different cost structure: less customer-facing labor no merchandising or presentation costs and higher delivery costs

For an analyst reading a grocer the useful question is what proportion of online volume is covered by each method and whether the mix is ​​moving as this determines the marginal profitability of online growth. A retailer that grows its online sales exclusively through in-store selection is developing a channel that may be losing money at the margin

That's the question worth hanging on to because online growth reads as unequivocally good in a headline and it isn't. Two grocers can report the same online growth rate and if one achieves it from sites converted to improve picking productivity while the other achieves it from commercial stores with pickers competing against buyers they are running different businesses with different marginal economics. The growth rate doesn't say which is which. The fulfillment mix does

The Bottom Line

Dark stores exist because online groceries have a pickup cost and a delivery cost that improve in opposite directions and the answer for most urban markets lies somewhere between a brick-and-mortar store and a distant automated warehouse. They convert low-performing retail properties into distribution capacity exactly where customers are which is the scarce resource. The version that failed was the one that promised ten-minute delivery from very small sites and it failed for the simple reason that the required density never arrived. The LessonIt generalizes beyond the grocery store: When a customer promise sets your service radius you've already set your maximum volume and all other decisions are made within a box you drew first