The STR Deal Underwriting Manual

The problem · Why most STR underwriting fails

Three errors account for most bad STR purchases, and all three happen before closing:

Optimism dressed as a projection. A market data tool gives you a number, the number goes straight into the model as "revenue," and the model produces a return that justifies the purchase. Nothing in that chain tested whether the number was achievable for this operator, in this first year, at this listing's real position in the market.

Confusing top-line collection with rental revenue. The revenue a comp tool reports can bundle cleaning fees the guest paid and, depending on the source, any occupancy taxes collected on the booking. Neither is yours to keep — cleaning is a pass-through cost, occupancy tax belongs to the taxing jurisdiction — yet both inflate the headline and quietly distort every ratio built on top of it.

An incomplete cost stack. The mortgage, taxes, and insurance are easy to remember. The dynamic-pricing software, the resupply, the turnover labor, the management fee, maintenance and repairs, the capital reserve, and the tens of thousands it takes to furnish the place before it earns a dollar are the lines that turn a "good deal" into a break-even one. The capital reserve is money set aside each year to replace things that wear out.

Underwrite the property you're buying, not the one you're hoping for. A defensible model uses conservative inputs and still clears your bar. If a deal only works on best-case assumptions, the assumptions are the product being sold — not the property.

Step 01 · Realistic assumptions

Every model rests on a handful of inputs: average daily rate (ADR), occupancy, seasonality, and the operating-cost ratios. ADR is the average nightly price of the nights actually booked; occupancy is the share of available nights that are booked. Get these right and the arithmetic downstream takes care of itself. Get them wrong and no amount of spreadsheet precision will save you.

Start with a real comp set (comparable listings used as evidence), not a single market average. Pull the closest matches — same bedroom count, same guest capacity, same sub-market and drive time, similar amenities and review depth — and throw out the outliers at both ends. The top listing in a market is usually run by a professional operator with a mature review history and a photography budget; underwriting to their number is underwriting to someone else's business.

Then clean the comp before you use it. Strip out cleaning-fee revenue and occupancy taxes so you are left with the true nightly rate a guest pays for lodging — the only figure that belongs in an ADR or RevPAN calculation, and the base every downstream ratio depends on.

The property in this manual is the example deal these guides use, an illustration rather than a forecast, whose full assumption set is listed in The Assumptions Behind Every Model. Its comps show a gross nightly rate around $340; stripped of the cleaning fee and pass-through taxes, the net ADR is about $285.

Finally, discount for the year you'll actually have, not the year the comp set already had. Market tools report what mature listings did; you are a new listing. Underwrite to a stabilized occupancy below the market's — roughly 62% against a market nearer 70% — for conservatism, then take the first year lower again for the ramp, rounding it down rather than up. Modeling that ramp is the next step.

Step 02 · Revenue, projected conservatively

Revenue is not a single number; it is average daily rate times occupancy across a seasonal year, discounted for the fact that a brand-new listing does not perform like an established one. The metric that matters most here is RevPAN — revenue per available night — because it combines rate and occupancy into one figure. That stops you from being fooled by a high rate at low occupancy, or the reverse.

Model the ramp, don't assume it away. A new listing earns fewer bookings and less pricing power in its first three to four months while it accumulates reviews and search position — and while the platform says its search is designed to help new listings show up well, that is not a cushion you can underwrite. A two-phase year — suppressed months 1–4, then stabilized — is far closer to reality than a flat market average applied from day one.

For the example property, a net ADR of $285 at 62% occupancy is a RevPAN of about $176.70 per available night; across 365 nights that is $64,495.50 — the stabilized-year lodging revenue that goes into the model, not the $85,000 the tool first returned. The first year runs lower while the listing ramps; rounded down, call it about $63,000.

RevPAN, not ADR, tells you whether a deal works. A headline nightly rate means nothing without the occupancy to fill it. Underwrite the combination, and discount it for the year the property will actually have — not the year the comp set already had.

For more detail on each piece, see RevPAN as the core metric and modeling the first-year ramp-up.

Step 03 · The real cost stack

Two categories, and operators forget different lines in each. Fixed costs continue whether or not the property books: property taxes, insurance (at the elevated short-term-rental rate, not a standard homeowner's policy), any homeowners-association (HOA) dues, and the base utilities and internet a listing must keep on. Variable costs scale with bookings: cleaning and turnover labor, consumables and resupply, platform service fees, and the dynamic-pricing and property-management software that a serious listing runs on.

Then come two lines that are technically not operating expenses but decide whether the deal is even fundable. Neither is counted in NOI (net operating income), which is what the property earns after operating costs and before loan payments.

The capital reserve. A capital replacement reserve is not optional. Furnishings wear out on a short cycle under heavy turnover, and the roof, HVAC, and appliances are yours. For the example it is $13,206 a year, set for this property, not a universal percentage, and it sits below NOI. Underwriting without a reserve line is underwriting a property that never ages.

Start-up capital. Before the property earns a dollar it has to be furnished and equipped — furniture, mattresses and linens, kitchen and operating supplies, design, professional photography, smart locks and noise monitoring. For our example that initial outlay is roughly $45,000. It is not an annual cost, but it is real cash out the door, and — critically — it belongs in the denominator of the return calculation in Step 05. Leave it out and the cash-on-cash return you compute is distorted, overstated whenever the deal earns cash.

For the example property, the illustrative operating stack before management — fixed costs $15,200, variable costs $6,300 and maintenance and repairs $2,500 — comes to $24,000; a full-service manager at 22% of revenue adds $14,189, for operating expenses of $38,189. The capital reserve is counted separately, below NOI. The seasonal cash reserve that sits alongside these costs is covered in Step 06.

Step 04 · Financing & capital structure

Financing decides two things: how much cash you actually put in, and how much of the property's income the lender takes back each year before you see any of it. Both feed the returns in the next step.

Total cash invested is more than the down payment. For the example, 25% down on $650,000 is $162,500. Add roughly $19,500 in closing costs and the $45,000 furnishing outlay, and the real capital at risk is about $227,000, not the $162,500 the naive version counts. On the loan side, $487,500 at roughly 7% on a 30-year term is about $38,920 a year in debt service. Debt service is a year's loan payments, principal and interest.

Leverage magnifies the outcome in both directions: it lowers the cash you commit and raises your return when the property performs, and it does the exact opposite when it doesn't. Whether to lean into it — and how the same deal looks in cash versus financed — is its own decision, covered in the Financing discipline and in Should I Pay Cash or Finance My STR?.

Step 05 · The returns that matter

Four numbers, each answering a different question. None of them is sufficient alone.

Cash-on-cash return — pre-tax cash flow after debt service and the capital reserve (NOI minus annual debt service minus the capital reserve) divided by total cash invested — is the one that most reflects your actual experience as the owner. It is also the one most often inflated, because the denominator quietly drops the furnishing and closing costs from total cash invested, and the numerator quietly uses the un-discounted revenue.

Cap rate — net operating income over purchase price — is useful for comparing properties, but a cap rate is only as comparable as the NOI beneath it. For an apples-to-apples read, use NOI after a management allowance — a market management cost included even if you self-manage — so on this deal the realistic cap rate is about 4.05% ($26,306 ÷ $650,000). On NOI before that allowance ($40,496) it would read 6.23%; that is the figure for a self-managed owner's own planning, not for comparing assets. Treat it as one lens, not the verdict.

Debt-service coverage ratio (DSCR) — net operating income divided by annual debt service — is the operating coverage check: whether the property's own income pays its loan. It is not a lender's qualification test. A DSCR lender runs its own qualifying math, which can differ from this ratio. Lender program minimums — commonly somewhere around 1.0 to 1.25, with 1.25 or higher commonly earning the best pricing (a market benchmark, not a rule) — apply to that qualifying ratio, not to this one. On this deal the operating DSCR is 0.68 ($26,306 ÷ $38,920): the property's income does not cover its own debt on realistic numbers.

Break-even occupancy — the share of nights you must book to cover operating costs (management included), the capital reserve and debt — tells you how much cushion you have. With management at 22%, this property breaks even at 98.4% against the 62% you are projecting: 36.4 points, about 133 nights a year, short. Even before a management allowance it needs 74.4%, 12.4 points (about 45 nights) above the projection. There is no cushion on either basis.

Here is the example property underwritten both ways — the same house, the same price, two sets of assumptions:

Each dollar figure is rounded independently, so the displayed amounts may not add up exactly. Stabilized revenue is $64,495.50 and operating expenses $38,189.01, so NOI is $26,306.49; that is why the rounded column appears to be $1 out.

The property did not change. The rate did not change. The only thing that changed was how realistic the inputs were, and an 18.5% "great deal" became one that loses about $25,820 a year after its reserve and does not cover its own debt. The full method behind two of these lines lives in the break-even occupancy calculator.

Step 06 · Downside protection & reserves

A single-line annual model hides the two ways short-term rentals actually run into trouble: the calendar and the bad year.

The seasonal cash reserve exists because STR income is uneven through the year. Peak months run well ahead of costs; slow months run behind them. The reserve is sized to cover the predictable gap between slow-season revenue and monthly fixed costs including debt. You fund it in peak months and draw on it in slow ones. It is an annual cycle, not a one-time buffer, and without it a property that breaks even on paper can still fail on cash timing.

The stress test asks what the return looks like when the year disappoints. Re-run the model with ADR down 15% and occupancy down 10 percentage points from your already-conservative base. A deal that survives that and still covers its debt has real margin; a deal that goes cash-flow negative under a mild stress was never as safe as the base case implied.

Step 07 · The verdict: go / no-go

Underwriting ends in a decision, not a spreadsheet. Assemble the realistic numbers and hold them against a bar you set before you fell in love with the property: a minimum cash-on-cash return, a DSCR comfortably above 1.00, and a break-even occupancy that sits well under what the market realistically delivers.

The example property fails all three at the asking price — cash flow of about −$25,820 a year after its reserve, a DSCR of 0.68, and a break-even occupancy of 98.4% against the 62% it projects (74.4% even before a management allowance). That is not a bad property; it is a No-Go at $650,000. If the terms change — a lower price, more equity in, a different loan — re-underwrite the deal on the new terms rather than assuming the verdict flips; a stronger market is a different deal — an alternative use of your capital, not a fix for this one. The discipline is being willing to walk when the realistic numbers say walk.

The Builders Finance Underwriting Method — the four moves, applied to every number:

• Source it — trace every number to real evidence, not a headline.
• Haircut it — discount for the year you'll actually have; round revenue down, costs up.
• Record it — value, source, and haircut, in The Assumptions Ledger.
• Stress it — move the numbers that matter to their downside before you trust them.

Each step has its own guide. This manual is the overview. Start with the guides to the individual steps — The Assumptions Behind Every Model, Estimating STR Revenue Without Fooling Yourself, What STRs Actually Cost to Run, Cash-on-Cash Return for an STR, and Cap Rate for STRs (and Why It Misleads) — then bring them together in the two decisions: Does This Deal Actually Pencil? scores a specific deal, and Should I Buy in This Market? screens the market before you underwrite.

When the question is whether a comp tool's number can be trusted, Is AirDNA (or Rabbu) Accurate? goes deeper. When a property you already own isn't producing what the model said, Why Isn't My Airbnb Cash Flowing? is the diagnostic. For the deeper metric treatments, see RevPAN and first-year ramp-up modeling, and run your own numbers through the Honest STR Underwriting Calculator and the Break-Even Occupancy Calculator.