Somewhere in your network this morning, a customer walked into a branch, handed over a prepaid parcel and left. A minute at the counter, maybe less. A few pence of margin, on a parcel somebody else had already priced and already sold.
Behind that counter there is a room. It has shelves in it. There is nothing on those shelves that belongs to a paying merchant.
That is the whole argument, and most postal strategies walk straight past it.
In the UK, the House of Commons Library counted 11,708 post office branches in February 2026, against a funded floor of 11,500 and access criteria requiring 99% of the population to be within three miles of a branch. Drop and collect branches, which handle prepaid parcels and bill payments rather than a full post office service, made up 7% of the network in March 2025. The Post Office is aiming for 3,000 of them, roughly a fifth of the estate. (House of Commons Library, Post office branches: statistics and developments, briefing SN02585)
The UK is the worked example here because its numbers are public, and because it is the hard case, for a reason that comes up later.
The network is being deliberately re-shaped around parcels. It is being re-shaped at the door.
Drop and collect is a handover point. The customer brings a parcel a marketplace or retailer has already priced, already labelled and already taken the margin on. The branch provides floor space, a member of staff and a scan. It is a real service. It is also the thinnest-margin square metre in the building.
Parcels are growing. Parcel money is not following.
FreightWaves reported this year that postal revenue growth has come almost entirely from parcels rather than mail, while margins in that category stay narrow or negative. Norway's Posten Bring is the clean illustration: first-half 2026 revenue up 0.9%, adjusted operating profit down 19.1%, parcel volumes still climbing. (FreightWaves, Postal operators struggle to break even despite parcel growth)
Meanwhile the value is being built a short drive away. GCC quick commerce is forecast to grow from around USD 4.6bn in 2026 to USD 12.4bn by 2031 (Mordor Intelligence, GCC Quick Commerce Market), and Redseer argues the GCC is the first region globally where dark store economics can be made to work (Redseer, Dark stores can be profitable, GCC is where it happens first).
Those operators are all spending money to buy proximity. Sites, leases, fit-out, pickers, routing. They are building, at considerable expense, a thinner version of something national posts were handed by statute and have carried on the cost line for decades.
Defend the counter and wait. Keep the estate for the universal service, grow drop and collect, hope volume eventually drags margin up with it. This is the default, and it is a bet that the last three years reverse.
Build purpose-built micro-fulfilment. New sites, new leases, new capital, competing where the specialists already have density and a head start.
Rent the network to somebody else. Let a marketplace or a quick commerce player use your branches as their nodes. Fast, low risk, and it makes you the landlord of your own advantage: someone else sets the price per order and keeps the merchant.
Use the estate you already have as fulfilment infrastructure. A post office branch becomes a fulfilment node: it holds third-party stock in a fraction of the space behind the counter, picks and packs it, and puts it into a round that already exists.
The fourth is the one nobody writes a strategy paper about, because it is not a procurement decision. It is an operating model decision, and those are harder to sponsor.
There is a familiar version of this argument that says any building with spare space can become a fulfilment centre. That undersells what a postal network actually has, and some of it cannot be bought.
Ubiquity nobody had to justify. No retailer, however large, has a board that would approve 11,500 sites covering 99% of the population within three miles. A post has them because it is required to. The most expensive part of a hyper-local fulfilment network is already built, already staffed and already paid for.
The round. In most markets the organisation holding the counter also drives to nearly every address in the country, every working day. Fulfilment and delivery sit inside one P&L. No warehouse software creates that.
Standing as a regulated institution. Handing stock to a national post is an easier internal conversation for a merchant than handing it to a two-year-old start-up.
The border. Posts already sit on cross-border rails, customs data obligations and duty handling. Local fulfilment stops being simple the moment an order originates or returns across a border, which is exactly where most local fulfilment propositions quietly fall over.
In the UK this is harder than elsewhere, because Post Office Limited was separated from Royal Mail in 2012. The counter and the round sit in different companies, so the advantage above has to be assembled commercially rather than assumed. In most other markets one operator holds both.
Most branches are also run by postmasters and franchise partners who are independent businesspeople. You cannot mandate a fulfilment operation into a network you do not employ. The adoption problem is a postmaster P&L problem, not a technology problem, and any version that does not put a visible new income line into that P&L will be politely ignored.
And holding somebody else's stock creates liability, shrinkage and space questions a prepaid parcel never does. That is a governance conversation before it is an operations one.
If an approach fails one of them, it is not ready.
That last one matters more than the technology. The operators who have moved on this are not the ones with the biggest budgets. They are the ones who defined what failure looked like in advance, and could therefore get approval to start. We set out how to structure that pilot, and what evidence a board will accept, in our guide to launching a dark-store pilot. The tactical layer underneath it is covered in an earlier piece: Dark Stores: The Revenue Model Hidden in Plain Sight.
Saudi Post Logistics had this problem at scale. Merchants across cosmetics, telecoms, fashion and SMEs, each on a different platform and stock system, and a network that could not take more without adding people.
First it pulled every merchant's stock, products and orders into one system, whatever they sold on. Then branches started holding that stock and picking orders, which was an extension of what already ran there, not a building project.
Fast-moving lines sit in branch dark stores. The long tail stays in a mega centre, and an order goes to whichever can serve it. The branch that picks the order is where the customer collects it, and where it lands if it comes back.
In the first half of 2025, against the same half of 2024, processed orders grew 140% and dark store volumes 17%, with no additional headcount for the expansion. The team could see stock, prices and orders while they were still moving.
SPL had the same legacy systems as everyone else. Same unions, same regulator. The branches were already there. They just had nothing on the shelves.
GN TEQ built that layer with SPL, and we build it with posts, forwarders and carriers elsewhere. It sits over the postal system, the TMS and the branch estate you already run: one place to hold one version of stock, run orders through branches, and see cost and margin per parcel while the week is still happening. We do not replace anything, and nobody has to take the counter apart.
If this matches a conversation you are already having internally, the useful next step is not a demo. It is picking the ten branches you would test on.
So: if a merchant offered your network 500 orders a week starting Monday, which branches would you give them, and could you tell by Friday whether you had made money on it?
Yes, and the expensive part is already built. A post office branch can hold third-party stock in a fraction of the space behind the counter, pick and pack it, and put it into a delivery round that already runs to nearly every address. What makes it work is not the shelving. It is having stock, product data, pricing and orders in one system rather than four.
Most postal revenue growth now comes from parcels rather than mail, but margins in that category stay narrow or negative. Norway's Posten Bring is the clean illustration: first-half 2026 revenue up 0.9%, adjusted operating profit down 19.1%, parcel volumes still climbing. Much of that growth arrives as prepaid drop and collect, where a marketplace or retailer has already priced the parcel and taken the margin.
Drop and collect is a handover point. The customer brings in a parcel somebody else has already priced and labelled, and the branch provides floor space, a member of staff and a scan. Fulfilment means the branch holds the merchant's stock and picks the order, so the post earns on the order rather than on the handover.
Six things: no new estate or leases, no net new headcount at branch level, one version of stock across every channel the merchant sells on, margin per order visible at branch level inside the week, cross-border and duty handled in the same flow, and written exit criteria before the pilot starts. If an approach fails one of them, it is not ready. The operators who have moved on this are the ones who defined what failure looked like in advance, which is what let them get approval to start.