Key points
- The faster the delivery promise, the closer the inventory sits to the doorway; the final stretch can exceed half of logistics cost.
- Without intervention, delivery vehicles in the hundred largest cities will grow by a third before 2030, about eleven extra minutes of travel a day.
- Barcelona banned new ghost supermarkets and required waiting space for couriers: planning had to invent a space for invisible work.
The last mile on the outskirts is the stretch from the warehouse to your door and the place where the convenience of digital commerce is paid for in city. You press a button and two hours later somebody rings the bell; the operation looks immaterial, but that parcel has been moving for days: it left a shed the size of several blocks beside a motorway thirty or forty kilometres away, passed through an intermediate warehouse and a neighbourhood unit with its shutter down, and travelled in a van double parked while the driver ran up two flights of stairs. The question is how much city it takes for a parcel to arrive in two hours, and the answer is a geography of speed organised in layers.
Three layers of the geography of speed: sheds, warehouses and ghost units
The first layer sits far out: enormous sheds where land is cheap and there is motorway access. The second moves closer: mid-sized warehouses on the metropolitan edge. The third is inside the neighbourhood, disguised in units that used to be shops and are now ghost supermarkets with no customers. The faster the delivery promise, the closer to the doorway the inventory must sit, and there lies the economic knot: the final stretch can account for more than half of total logistics cost, so all the pressure lands on the street. That model consumes land at a scale almost nobody notices: in Spain, logistics space take-up in Madrid and Barcelona alone passed a million and a half square metres in a recent year, well above the average of the previous decade.
Land, traffic and eleven minutes a day: the measured costs
That land competes with productive industry and, in peripheral municipalities, with housing, and its urban yield is low: a logistics shed occupies an enormous footprint, generates little employment per hectare compared with other activities and produces constant heavy traffic; municipalities accept it for the quick revenue while the environmental and road costs are shared among everyone. On the street the effects have been measured: a World Economic Forum analysis estimated that, without intervention, delivery vehicles in the world's hundred largest cities will grow by around a third before 2030, delivery emissions will rise by close to a third and congestion by somewhat more than a fifth, about eleven extra minutes of travel a day for every person. The time you save by not going to the shop is taken from whoever is driving down that street.
The labour layer: algorithms, pavements and Barcelona's rule
The third layer is labour. The promised speed rests on fragmented shifts, routes assigned by algorithm and times calculated to the minute; the courier does not control the pace, an application measuring every stop does, and many delivery units have no interior space to wait, so the waiting happens on the pavement with mopeds lined up. Barcelona banned in 2021 the opening of new ghost supermarkets across the whole city and required other premises to set aside waiting space for couriers proportional to their floor area, a revealing detail: planning rules had to invent a space for work the model assumed was invisible.
Facts, interpretation and counterargument: which rules make the last mile pay
It helps to separate facts from interpretation. The facts are that digital commerce consumes peripheral land, increases light traffic and rests on high-turnover employment; the interpretation is that we have shifted onto the city the costs of a service we perceive as free, because its price includes neither congestion, nor the night noise of unloading, nor road wear, nor the occupied pavement. Instant convenience works because its bill is paid in kind, with public space. The counterargument is solid: one van delivering fifty orders on an optimised route replaces dozens of private cars, several studies suggest that well organised delivery reduces total kilometres compared with shopping by car, and home delivery is an accessibility service for older people, people with disabilities or people without time.
The conclusion is that the problem is not delivery but its speed and its price. A city can require the last mile to pay what it costs: neighbourhood micro-consolidation centres that group the shipments of every operator, delivery by cargo bike and on foot, time windows and charges for using loading bays, limits on ghost supermarkets and decent spaces for those who deliver. With those rules the parcel still arrives; without them the periphery fills with sheds, the street with vans and the pavement with people waiting for an order that somebody, somewhere else, believed was free.
Frequently asked questions
How much city does a two-hour delivery need?
A geography of speed in layers: enormous sheds beside motorways where land is cheap, mid-sized warehouses on the metropolitan edge and ghost supermarkets inside the neighbourhood, plus double-parked vans, algorithmic routes and couriers waiting on the pavement; the final stretch can account for more than half of total logistics cost.
Is home delivery not more efficient than driving to the shop?
It can be: one van on an optimised route replaces dozens of private cars and delivery is an accessibility service for older or disabled people. The advantage disappears when firms compete to deliver within two hours, because vans leave half empty and repeat journeys, so the answer is to make the last mile pay its cost.