How to Underwrite a Self-Storage Development (Without Fooling Yourself)

A facility can have excellent rents and still lose money. Here's the chain of math that tells you which — before you spend a dollar on architecture.

A few years ago I underwrote a ground-up self-storage deal that looked like a winner. Multi-story, climate-controlled, dense infill lot in a high-cost Northeast metro. The rents were spectacular — a blended $3.57 per square foot per month, several times the national average, because urban storage commands a premium people happily pay. Stabilized net operating income penciled to about $442,000.

It lost money. Not a little — the model spat back a negative profit and an internal rate of return hovering around zero. And the thing that told me, in about five minutes, was a single number most first-time developers never calculate.

This guide is that number, and the chain of assumptions that produces it. By the end you'll be able to take a raw site — a price, a lot size, a market — and tell whether it's a deal or a trap, using the same sequence institutional shops run. No software required to follow along; a free calculator at the end will do the arithmetic.

Development is not a real estate business. It's a spread business. You build to one yield and sell at another, and the gap between them is the entire game.

0. The number that decides everything

That number is the development spread: your stabilized yield on cost minus the cap rate you'll exit at.

Yield on cost is stabilized NOI divided by what the whole project cost you to build. The exit cap is what a buyer will pay for that NOI when you're done — expressed as a yield, so a 6% cap means a buyer pays about 16.7× the annual NOI. You build to, say, an 8% yield on cost and sell at a 6% cap, and that 200-basis-point gap is your profit and your margin of safety against everything that can go wrong: cost overruns, a slow lease-up, cap rates drifting up while you build.

The widely cited threshold for a financially sound merchant-build is a spread of at least 200 basis points over the exit cap. With construction costs where they are in 2025–26, plenty of developers now want 250–300. Below ~150, you are taking development risk for a trading-level return — building for months to earn what you could have bought on day one.

My infill deal? Yield on cost came out to roughly 5.1% against an exit cap of 4.5%. A 60-basis-point spread. There was no room. No amount of optimism about those gorgeous rents could fix it, because — and this is the lesson the whole guide circles back to — storage is a low-rent-density use, and it cannot carry an expensive land basis. Land was 76% of total cost. Game over before the first comment from the planning department.

Everything below is how you get to that number honestly. We'll build one clean example as we go — a 55,000-square-foot single-story drive-up facility in a solid secondary market — and watch it earn its verdict.

1. Start with the box: gross area, efficiency, and rentable square feet

Storage is sold by the square foot, so everything starts with how many rentable square feet you can squeeze out of the building you're allowed to build. Two numbers matter:

Gross building area (GBA) is the whole footprint. Net rentable square feet (NRSF) is what actually earns rent — GBA minus hallways, stairs, elevators, the office, mechanical rooms, and wall thickness. The ratio between them, the efficiency ratio, is the first place rookies quietly destroy a pro forma by assuming they'll rent space they can never rent.

• Single-story drive-up: 80–88% efficient. Little is lost to circulation.
• Multi-story, elevator-served: 65–75%. Elevators, structural framing, ADA corridors, and a tighter envelope eat your rentable area.
• The non-obvious one — ramp / over-under designs: ~88–90%, nearly single-story efficiency, because they eliminate elevators and the dead space around them.

That spread is enormous. On a 100,000-square-foot building, the difference between 70% and 85% efficiency is 15,000 rentable square feet you either collect rent on for thirty years or pay to build and heat for free. Multi-story isn't wrong — it's how you make dense, expensive land work — but you pay for vertical in lost efficiency, and the pro forma has to absorb it.

2. Cost the build — and respect 2025 reality

Construction cost per square foot is the most-Googled number in this business and the most badly used. The trap: published figures are almost always vertical construction — the building package and erection — not all-in, which adds land, site work, soft costs, and contingency. Confuse the two and you'll underwrite a deal that's 40% light on cost.

Current ranges, per Storable's late-2025 cost guide and contractor data:

*All-in figures are syntheses and swing hard on land price; treat as ranges, not gospel.

Two 2025–26 cost drivers belong in your model that weren't there two years ago. Light-gauge steel framing is up 9–14% year over year on tariffs and tight mill schedules. And on climate-controlled product, the EPA's A2L refrigerant transition is adding roughly 6–10% to mechanical cost. If your cost basis came from a 2023 spreadsheet, it's wrong.

Build your total project cost the way a lender will: land, then hard costs (the building), then site work, then soft costs (architecture, engineering, permits, impact fees, financing — call it 12–18% of hard), then a contingency of 7–10% because something always comes up, then a developer fee. Add construction-loan interest and you have your all-in number — the denominator of yield on cost.

3. Build the revenue — and don't blend what shouldn't be blended

Here's where storage is genuinely different from an apartment building, and where the cleanest underwriting mistakes happen. Smaller units rent for dramatically more per square foot than larger ones. A 5×5 might fetch $1.40/sf/month while a 10×20 fetches $0.70 — the small unit earns double the rent per foot. Your unit mix, not some single market rent, drives your blended rate. Underwrite one average $/sf across the whole building and you'll be off by 20–30% depending on how the mix actually shakes out.

So you build a mix table: how many of each size, at what monthly rate, and let the rentable square feet reconcile back to the NRSF from Step 1. Then layer in the income first-timers forget entirely:

Tenant insurance — the quiet 8–10% of revenue

Nearly every well-run facility sells tenant insurance or a tenant-protection plan at move-in — $9–13/month for a couple thousand dollars of coverage. Penetration runs 50–60% when it's voluntary and 80%+ when the lease requires proof of insurance or enrollment. The underlying risk costs the operator a fraction of that, so margins are high, and at scale it contributes on the order of 8–10% of total facility revenue. (Vendors quote ~89% margins; discount that enthusiasm, but the line item is real and it's pure leverage on NOI.) Leave it out of your pro forma and you've understated value by more than the line itself — you've understated the NOI that gets capitalized.

Then admin and late fees, retail (locks, boxes), and maybe truck rental. None is huge alone; together they're the difference between a thin deal and a good one.

4. Be honest about lease-up — this is where pro formas lie

A new facility does not open full. It opens empty, and it fills slowly, and the single most common way a first-time storage pro forma becomes fiction is assuming stabilized occupancy in year one.

Real absorption runs roughly 2–3% of rentable space per month — a useful rule of thumb is 1,200–1,500 net rentable feet leased monthly, then calibrated to your feasibility study. Time to stabilization (≈85–90% occupied) has actually lengthened: the planning figure has crept from 30 months to 36 months as facilities got bigger and new supply got heavier. A large ground-up project taking two to three years to stabilize is normal, not a red flag.

This matters for two reasons. First, your returns are driven by when the cash shows up, not just how much — a deal that stabilizes in 36 months instead of 18 can see its IRR cut in half even with identical stabilized NOI. Second, you have to fund the gap: an interest reserve or operating shortfall reserve that carries the debt service while the building fills. Forget that and you'll run out of money in month 14 of a 36-month lease-up, which is how good projects die.

You underwrite this with a monthly curve — occupancy climbing month by month toward stabilization — and you let that curve, not a year-one snapshot, drive the cash flows behind your IRR.

5. Operate it — and don't trust the seller's expenses

Self-storage runs lean. Operating expenses land around 35% of effective gross income nationally (the 2024 expense guidebook puts it at 34.7%); the big REITs, with scale, run 26–31%. The line items: property taxes (usually the single largest, and they'll get reassessed on your new building — underwrite the future bill, not the raw-land bill), a management fee of 5–6% of revenue, payroll, marketing, repairs, insurance, utilities, and software.

6. The verdict

Now the chain closes. We have a cost and we have an NOI, so we have a yield on cost. We pick an exit cap from the market, and we have our spread.

Two hundred basis points of spread, a third of cost in profit. This one you underwrite the rest of the way — order the feasibility study, price the dirt, take it to a lender. The chain held.

Now put my infill deal through the same gate. Elite rents, $442K of NOI — but $8.6M of cost, because land alone was $6.55M. Yield on cost 5.1%, exit cap 4.5%, spread 60 basis points. Same math, opposite verdict: walk. The rents were never the problem. The basis was. Cap rates being what they are in coastal metros — high-4s to mid-5s for stabilized product — there simply wasn't a spread to be had on land priced for apartments.

Strong rents don't save a bad basis. Storage can't carry expensive dirt — and the spread test catches it in five minutes.

The five mistakes that sink first-time storage developers

1. Skipping the spread test. Run yield-on-cost minus exit cap first, before you fall in love. Everything else is detail.
2. Assuming you open full. Model the 24–36 month lease-up and fund the gap, or you'll run out of cash mid-fill.
3. Blending one rent across all units. A 5×5 earns double the $/sf of a 10×20. The mix is the model.
4. Trusting raw or seller expenses. Add a real management fee, payroll, insurance, and reassessed taxes. NOI is usually overstated 10–15%.
5. Paying an apartment price for storage dirt. Storage is low-rent-density. If land is most of your cost, the spread won't be there — no matter how good the rents look.

Where to go from here

Go deeper on each link in the chain: the two yields that decide the deal in cap rates & yield on cost; the all-in construction cost per square foot; the unit mix that sets your blended rate; the lease-up curve that drives your IRR; the development pro forma that strings it across time; and the feasibility study that grounds all of it in a real market.

Underwrite the deal, not the dream. If you want the arithmetic done for you, the free calculator gives you yield on cost and a go/walk verdict from a handful of inputs. When you're ready to underwrite for real — unit-mix revenue, the full monthly lease-up curve, profit, equity multiple, and IRR, with every assumption explained — the complete Excel model is the same one I run my own deals through, and it ships with a full teardown of the infill deal above.

Useful third-party resources: StorTrack for submarket rate and supply data; Stora or other management platforms for operations once you're built. This article is educational, not investment advice — every deal turns on its own diligence.