Estimating STR Revenue Without Fooling Yourself

Why revenue is where deals die

Revenue is the easiest number in an underwriting model to overestimate, and it is the one every other number depends on — so a mistake here doesn't stay here.

Overstate revenue and your cash-on-cash return (a year's cash flow as a share of the total cash you put in), your debt coverage and your break-even occupancy all inherit the error. And because the mistake sits at the top of the model, it doesn't stay in proportion: a 15% revenue miss doesn't shave 15% off the return. On a leveraged deal (one financed with a loan), the miss can erase the return. The good news: revenue is not a mystery you have to guess at. It is a build with a small number of steps, and each step has a realistic way and an optimistic way to do it. This guide walks the realistic way.

(For the discipline behind treating every input this way — sourcing it, discounting it, writing it down — see The Assumptions Behind Every Model.)

Step 1 — Build a real comp set

Everything downstream rests on the comparables you choose, so choose them like an appraiser, not like an optimist. You want eight to twelve active listings that genuinely match the property: same bedroom and bathroom count, same guest capacity, same sub-market and drive time, and similar amenities that actually move bookings — a hot tub, a pool, a view, walkability. Then drop the outliers at both ends.

Two comps deserve special suspicion. The trophy listing at the top of the market is usually run by a professional operator with a mature review history and a real photography budget; underwriting to their number is underwriting to someone else's business. And a brand-new listing with three reviews tells you nothing yet. The middle of a clean comp set is worth more than either end. (How far to trust the tools you pull these from — AirDNA, Rabbu, and the rest — is covered in Is AirDNA (or Rabbu) Accurate?.)

Step 2 — Estimate the net rate, not the gross

Average daily rate is the price a guest pays for the room — and the number you pull from a comp or a tool usually isn't that. Reported "average nightly rate" can bundle the cleaning fee the guest paid and, depending on the source, any occupancy taxes collected on the booking. Neither is revenue you keep: cleaning is a pass-through cost, and occupancy tax belongs to the taxing jurisdiction. Leave them in and every ratio built on the rate is inflated from the first step.

So strip them out to get net ADR — the lodging rate alone — and take the median of your cleaned comps rather than the mean, so one luxury outlier doesn't drag the estimate up. Then adjust once, realistically, for how your specific property compares to the comp median: a dated interior and no hot tub earns a rate below the median, not at it.

Take the example deal these guides use, an illustration rather than a forecast, whose full assumption set is listed in The Assumptions Behind Every Model: a comp median gross rate of around $340 becomes a net ADR of $285 once the cleaning and tax pass-throughs come out.

Step 3 — Estimate your occupancy, not the market's

Occupancy is booked nights divided by available nights, and it is the single most over-read number in the estimate. The market average you'll be handed is too high for your listing for two reasons, and it's worth knowing them by name: survivorship (the average only includes listings that made it, not the ones that quietly failed) and blocked nights (tools measure occupancy against available nights, so an owner who blocks the slow season looks fuller than they are). A third, the ramp (a new listing books less in its first three to four months while it earns reviews and search position), is real but belongs to the first year only, so it is modelled as a separate step, not taken out of the occupancy.

The move is the same regardless: start from the market number and take it down to the occupancy your listing can hold once it has stabilized. In the running example, a market occupancy near 70% becomes a stabilized underwriting occupancy of about 62% — roughly 226 booked nights out of 365. (Year one runs lower still while the listing ramps: The First-Year Ramp-Up (and Why Year One Isn't Your Stabilized Year) works through how to discount market data for a listing with no history of its own.)

Step 4 — Combine as RevPAN, then shape the year

Do not multiply your best-case rate by your best-case occupancy. Combine them as RevPAN — revenue per available night, which is ADR times occupancy — the realistic way to hold both at once. A high rate at low occupancy and a low rate at high occupancy can produce the same RevPAN, which is exactly the point: it stops you from stacking the optimistic end of one number on top of the optimistic end of the other. (RevPAN is the figure that makes those comparisons fair, and this guide builds it from the same inputs below.)

Then shape it across the season rather than leaving it as a flat annual average. Model it by month, or at least by peak / shoulder / off-peak (the shoulder months sit on either side of the busy season), because the average hides the reality that peak months carry the year while slow months run below cost. That shape is what tells you how large a cash reserve the property needs to survive its own calendar. (How to size that buffer is covered in What STRs Actually Cost to Run.)

Step 5 — The last step isn't math. It's judgment.

Never carry the market's number into the model. Carry the haircut one, and keep the first year separate. The build produced $64,495.50 a year — $64,496 — at an occupancy already taken down from the market's: that is the stabilized revenue you underwrite to. The first year runs lower while the listing earns its reviews and search position; in the running example call it about $63,000, and use that for any first-year cash projection. If the deal only works on the market's number rather than the haircut one, it doesn't work.

Before you trust it, move it

Even a carefully built revenue number is one number, and one number is fragile. Run occupancy down another eight to ten percentage points and the rate down another ten to fifteen percent, and see whether the deal still covers its debt. Revenue is the input the whole model is most sensitive to, so if a plausible soft year takes the deal underwater, that isn't a detail — that's the finding. (The occupancy a property has to hold just to cover its costs is the other half of this number: the break-even occupancy calculator works it out from your own figures.)

Moving the number like this is the last of the four moves in the Builders Finance Underwriting Method; The STR Deal Underwriting Manual covers all four.