Margins Power

Self-storage: the business that sells you space you already own

30 September 2026

A shed, a fence and a payment link. Why the margins are absurd, why the payback is slow, and the one number that decides whether a site works.

£11,021Revenue at 82% full
78.0%Gross margin
38 monthsPayback
4Owner's week

Self-storage is the closest thing in the physical world to software margins. You buy or lease a shell, divide it with sheet metal, put a lock on each door and a card reader on the gate. After that the cost of the two hundredth customer is the same as the cost of the first: nothing. There is no stock, no delivery, no staff on site, and the customer pays monthly, in advance, indefinitely.

The product is not storage. The product is the customer's unwillingness to decide what to throw away.

Where the money comes from

Revenue is four numbers multiplied together, and nothing else: how many units you built, how full they are, what you charge per square metre, and how big the average unit is.

revenue = units × occupancy × rate × size
Everything else in the business is a cost line.

At 120 units averaging 8.0 m², charging £14 per m² per month, at 82.0% occupancy, the site bills £11,021 a month. That is the entire top line. There is no cross-sell, no upsell worth modelling, and no seasonality beyond a mild summer bump when people move house.

The model

Move the sliders. The verdict line at the bottom uses the rule most operators actually apply: a site is worth building if it returns the fit-out inside four years at a realistic occupancy.

Site calculator

£11,021Monthly revenue
£5,521Monthly costs
£5,500Net per month
94.0%Gross margin
26 monthsPayback
45.5%Return on capital
Worth building. The fit-out comes back inside four years.

Two things fall out of the model that are not obvious until you move the sliders. First, price matters more than size: adding 20% to the rate adds 20% to the net, while adding 20% more units adds only about 14% because the fit-out and the rent go up with them. Second, occupancy is the whole business — below roughly 60% almost every site is loss-making, and above 85% almost every site prints money. There is very little in between.

Where the costs actually sit

Rent 38.1%Utilities 3.4%Admin and software 2.6%Marketing 6.0%Net 49.9%

Notice what is not on that chart: staff. A 120-unit site with electronic access does not need anyone on it. The owner's four hours a week are spent chasing two late payers and letting one new customer in. This is the single reason the business is attractive and the single reason it is priced the way it is — everyone can see it.

The fit-out, honestly

LineTypicalNote
Partitioning, 120 units€78,000€650/unit for sheet steel; more for climate control
Doors and locks€24,000€200/unit, roller shutters
Access control and gate€18,000Keypad, cameras, the software that actually matters
Electrics, lighting, fire€21,000Where the surprises live — see below
Flooring and paint€9,000
Signage and first fit€6,500
Permits, drawings, legal€7,500Assumes the use class is already right
Working capital to break-even€31,00018 months of shortfall at a realistic fill curve

Fire compliance is the line that moves. Retrofitting sprinklers into a shell that does not have them adds €40k–€90k, and whether you need them depends on what tenants are allowed to store — which is a decision you make, not one made for you.

Operator interviews, Spain and Portugal, 2026

The fill curve, which is the whole risk

Every plan assumes a fill curve and almost every plan is wrong about it. Here is what the real one looks like, why month nine is where operators quit, what the marketing actually costs per unit let, and the three location tests that predict occupancy better than any demographic study.

The eighteen months that decide it

Break-even occupancy36.2%
Units needed to break even43.4
Cash burned before break-even£31,000

The break-even line above is the number to memorise. At the default settings a site needs 0.0 of its 120 units let before it stops losing money. Most operators reach that somewhere in month eight to eleven. The ones who fail are the ones who financed the fit-out on an eighteen-month bullet.

Three tests for a site

  1. Drive time, not distance. Demand collapses beyond twelve minutes. A site two kilometres away across a river is a different market from one two kilometres away on the same road.
  2. Flats above shops within that drive time. Storage demand is a function of housing without storage. Count balconies, not people.
  3. Existing competition is a good sign, not a bad one. An area with two full sites has proven demand. An area with none usually has none for a reason.
I spent four months on a demographic study and the thing that predicted it was whether the flats nearby had garages.
Operatortwo sites, Porto

How it is sold

A stabilised site trades at six to nine times net, and the buyer is usually either a regional operator rolling up or a property investor who wants the yield. The multiple is set almost entirely by lease length: a freehold or a twenty-year lease sells at nine, a five-year lease sells at six, for the same profit. If you intend to exit, the lease negotiation is worth more than three years of operating improvements.


Next in this field: the climate-controlled variant, which doubles the fit-out and the rate, and whether it is worth it.

Filed in the atlas