The seven-day test sounds like a rule of thumb. It is arithmetic — a specific division problem with a specific answer, and you either have the data to run it or you don’t.
Short-term rentals get their favorable treatment from an exception in the passive activity rules: an activity whose average period of customer use is seven days or fewer is not treated as a rental activity at all. Everything downstream depends on clearing that threshold. And yet most owners never actually calculate it. They know their place is “a short-term rental,” they assume the average is obviously under a week, and they find out otherwise at the worst possible moment.
Here is how to run the number, where the data comes from, and what quietly breaks it.
The formula
Average period of customer use = total rental days ÷ number of separate guest stays.
That’s it. Two inputs, one division. The difficulty is never the math — it’s getting both inputs right.
Total rental days means days the property was actually rented to guests. Not days it was listed. Not days it was available. Not days you blocked the calendar for yourself. Vacancy is invisible to this calculation in both directions — it neither helps nor hurts.
Number of separate stays means distinct periods of customer use — generally one per reservation, one guest party at a time.
A worked example
Take a cabin with a normal-looking year:
| Booking pattern | Stays | Nights |
|---|---|---|
| Weekend bookings (2 nights) | 28 | 56 |
| Week-long summer bookings (7 nights) | 11 | 77 |
| Midweek short stays (3 nights) | 9 | 27 |
| Total | 48 | 160 |
160 ÷ 48 = 3.33 days. Comfortably inside the threshold, with a lot of room to spare.
Now change exactly one thing. In February, a family relocating to the area books the cabin for 60 nights while they house-hunt. It’s one reservation, it’s easy money, and it feels like a win:
| Booking pattern | Stays | Nights |
|---|---|---|
| Everything above | 48 | 160 |
| One relocation booking | 1 | 60 |
| Total | 49 | 220 |
220 ÷ 49 = 4.49 days. Still fine — but notice how much moved. One booking, out of forty-nine, shifted the average by more than a full day.
Now imagine a quieter property: 20 stays, 70 nights, averaging 3.5 days. Add that same 60-night booking and you get 130 ÷ 21 = 6.19 days. Still under seven, but the margin is nearly gone. One more long booking and the year is lost.
That is the entire risk in this calculation. It is an average, so a single long stay carries the weight of many short ones — and the fewer bookings you have, the more damage one of them can do.
Where the data actually comes from
This is where most owners get stuck, and it’s a reporting problem rather than a tax one.
The earnings or payout report will not answer this question. It is organized around money and payout dates: gross, service fee, net, when it was sent. A single payout routinely covers several reservations, and payouts follow the disbursement schedule rather than the stay. Nothing in it reliably tells you how many separate guest stays occurred or how many nights each ran.
What you need is the reservation-level export — one row per booking, with check-in date, check-out date, and nights. Both major platforms provide it; on Airbnb it lives with your reservations rather than your earnings, and on Vrbo it comes out of the reservation history. Pull it, and the calculation is a sum and a count.
Two practical notes. Export monthly, not annually — platforms limit how far back you can reach, and a year-end scramble is how data gets lost. And keep the file. The export is the evidence behind the number; a spreadsheet with a total in it is not the same thing as the underlying record.
The details that change the answer
Direct bookings count too. If you take reservations off-platform — repeat guests, your own site, word of mouth — those stays belong in the calculation. Owners who run a mix frequently compute the average from platform data alone and get a number that isn’t theirs.
Multiple platforms need to be combined. Airbnb, Vrbo, and direct are one property, one activity, one average. Three separate reports, one calculation.
It’s per activity, and grouping changes it. Each property is generally its own activity unless a grouping election applies. Grouping combines the numerator and denominator across properties — which can rescue a marginal property or drag down a strong one. That’s a real decision with real consequences, and it interacts with your material participation hours too. Make it deliberately with your tax advisor.
Your own use isn’t a rental day. Time you or your family spend in the property is not a guest stay and its nights are not rental days. It has its own consequences elsewhere in the rules, but here it simply doesn’t appear.
Extensions and edge cases are worth asking about. A guest who extends, a booking that spans the year end, a stay cancelled partway through — these are the kind of details that are usually straightforward but occasionally aren’t. Flag them rather than assuming.
Track it during the year — this is the one number you can still influence
Almost everything else in this area is determined by the time you file. Average stay is different: it’s built from decisions you make one booking at a time, all year.
Run the calculation monthly. It takes a few minutes once the export habit exists, and it turns a year-end verdict into a live number you can steer:
- Under 4 days by mid-year — plenty of headroom. A long booking is a business decision, not a risk.
- Between 5 and 6 — the margin is thinning. Worth knowing before you accept a monthly stay.
- Above 6.5 — treat further long bookings as a decision with a tax cost attached, and get advice on where you stand.
An owner running this number in September has options. An owner who finds out in March has a fact.
Then the books have to match
The calculation and the ledger are the same story told twice, and they have to agree. If your books record gross bookings and per-property income properly, your revenue will reconcile to the reservation data that produced your average. If they record net payouts in one lump, the two will never tie — and the first person to notice will be whoever is asking you to substantiate the position.
Per-property books, gross recorded as gross, reconciled monthly. It’s the same discipline that makes the average-stay number trustworthy in the first place.
This article is general information about calculation and recordkeeping, not tax advice, and it is not a determination that any strategy applies to your situation. The rules contain nuances this article doesn’t cover — talk your specific facts through with your tax advisor.
Common questions
Do vacant days count against my average?
No. The calculation uses days actually rented divided by the number of stays. Vacancy affects your income, not this number.
What if I rent on more than one platform?
Combine them. One property is one activity — total the rental days and the stay counts across Airbnb, Vrbo, and any direct bookings before dividing.
Does one long booking really matter that much?
Yes, and more than people expect. It’s an average, so a 60-night stay counts once in the denominator but adds 60 to the numerator. On a property with few bookings it can move the result by two or three days on its own.
Which report should I be exporting?
The reservation-level one — one row per booking with check-in, check-out, and nights. The earnings or payout report is organized around money and won’t give you stay counts.
How long should I keep the exports?
As long as the related return is open to examination. The export is the evidence behind the number, so keeping only the calculated result defeats the purpose.
If you’d rather this just be handled
We keep the books for short-term rental owners — per-property, gross recorded correctly, reconciled monthly, with the reservation data kept alongside the ledger so the numbers behind your tax position hold together.
Get in touch for a free review of your books, or read why the short-term rental loophole is a records problem, why your Airbnb payout is not your income, and short-term rental bookkeeping basics. Nearby? See Atlanta short-term rental bookkeeping.
Recent Comments