Every short-term rental purchase rests on a revenue number, and that number is almost always produced by one of three methods with very different reliability. Knowing which one you are looking at is the difference between underwriting and hoping.
The three methods
Platform and data provider estimates. Modeled figures generated from market data. Useful for screening at the market level, unreliable at the property level, because the model does not know that the property has no view, sits on a switchback road, or has four bedrooms but only sleeps six comfortably.
Seller provided history. The most valuable data source and the one most often presented selectively. A seller showing you strong months is not lying, they are marketing. Ask for the full trailing twelve months from the platform dashboards themselves, not a summary spreadsheet.
Comparable based modeling. Building a revenue estimate from the actual observed performance of genuinely similar properties in the same submarket. Slower, and the only method we underwrite on.
What makes a comparable actually comparable
- Bedroom and sleeping capacity match. The most important variable in most markets, and one where being off by one changes the guest segment entirely.
- Same submarket, not same city. A cabin ten minutes from the parkway and one thirty minutes out are different products.
- Same amenity tier. Hot tub, pool, game room, and view quality. A comparable without your amenities overstates your position and one with amenities you lack understates the gap.
- Similar review count and age. An established listing with three hundred reviews outperforms an identical new listing. That is a real advantage the new property does not have in year one.
- Full twelve months of data, including the ugly months.
The year one discount
A new listing has no reviews, no ranking history, and no repeat guests. Underwriting a first year at stabilized comparable performance is the single most common modeling error we see. We model year one below the comparable set and let the property earn its way up.
We model revenue from comparables, not estimates
Every deal we bring to a client is built from actual booking data on genuinely comparable inventory, with the assumptions visible.
Apply NowThree tests before believing a number
The off season test. Ask for February, or whatever the trough month is in that market. Peak months tell you the ceiling. The trough tells you whether you can hold the property through a bad year. See cash reserves and seasonality.
The rate versus occupancy decomposition. A property hitting a revenue target through high occupancy at a low rate is a different business from one hitting it through strong rate at moderate occupancy. The second has fewer turnovers, less wear, and lower cleaning costs. See occupancy rates explained.
The supply trend test. Historical revenue is backward looking. If the submarket has added inventory faster than demand has grown, last year's numbers describe a market that no longer exists. Count active listings against the prior year before trusting any projection.
The expense side is where models actually fail
Revenue estimates are usually within a reasonable range. Expense estimates are frequently fiction, because they are modeled as percentages rather than built from quotes.
- Insurance quoted for the actual property and use, not estimated at a national percentage. In coastal markets this single line has ended more deals than any other. See the insurance guide.
- Property tax at the reassessed value after sale, not the seller's current bill.
- Cleaning as turnovers times cost, not a percentage of revenue.
- Maintenance and replacement at short-term rental wear rates, which are faster than long term rental rates.
- Management or co-host fees at the actual local rate, with the participation implications understood. See co-hosting versus self managing.
A model with a credible revenue line and an invented expense line is not a model. See how to analyze an Airbnb deal for the full build.
My BnB Accelerator, LLC is a real estate acquisition firm, not a CPA firm, and nothing here is tax advice. Our tax partner is AE Tax Advisors, an independent firm.
Keep reading
Frequently asked questions
Are Airbnb revenue estimates from data tools accurate?
They are useful for screening at the market level and unreliable at the property level, because the model cannot see view quality, road access, or whether a four bedroom property comfortably sleeps six or ten. Comparable based modeling using actual observed performance of similar properties is the more defensible method.
What makes a good short-term rental comparable?
A match on bedroom count and real sleeping capacity, the same submarket rather than the same city, the same amenity tier including hot tub, pool, game room, and view, a similar review count and listing age, and a full twelve months of data including the trough months.
Should I expect first year revenue to match comparable properties?
No. A new listing has no reviews, no ranking history, and no repeat guests, so underwriting year one at stabilized comparable performance is one of the most common modeling errors. Model year one below the comparable set and let the property earn its way up.
Why do short-term rental pro formas fail?
Usually on the expense side rather than the revenue side. Insurance estimated as a percentage instead of quoted, property tax taken from the seller's bill rather than the reassessed value, cleaning modeled as a percentage of revenue instead of turnovers times cost, and maintenance modeled at long term rental wear rates.