Occupancy, ADR, or RevPAR — Which One Should You Chase?
Work the 100-rooms-at-$150 versus 50-rooms-at-$200 argument through, and see why RevPAR wins the debate but still misses what the booking cost you.
Chase RevPAR, because occupancy and ADR can each be moved on their own without making you any money — and RevPAR cannot. But RevPAR is a rooms-only number. It does not know what you paid to get the booking, so once you are choosing between two roughly equal RevPAR outcomes, the deciding number is NRevPAR or GOPPAR. Here is the arithmetic, run all the way to the bottom line.
Settle the argument with arithmetic
The question every operator has argued about: 100 rooms at $150, or 50 rooms at $200?
Take a 100-room hotel for one night. Scenario A sells all 100 at $150. Scenario B sells 50 at $200. Scenario C sells 80 at $190.
| Rooms sold | Occupancy | ADR | Room revenue | RevPAR | |
|---|---|---|---|---|---|
| A — fill it | 100 | 100% | $150 | $15,000 | $150.00 |
| B — hold rate | 50 | 50% | $200 | $10,000 | $100.00 |
| C — middle | 80 | 80% | $190 | $15,200 | $152.00 |
Scenario B wins on ADR and loses badly. Scenario A wins on occupancy. Scenario C wins on RevPAR by two dollars — near enough a tie with A on the rooms-only view.
Your hotel is smaller than 100 rooms. The ratios do not care.
Why occupancy on its own means nothing
Occupancy is rooms sold divided by rooms available. It is the easiest metric in the building to move, because you control the price and demand for cheap rooms is effectively unlimited.
Any hotel can run 100% occupancy tomorrow. Price at $19 and you will. The number will look excellent in a report and the business will be worse off.
Occupancy is a diagnostic, not a target. It tells you how a night went. It never tells you whether the night went well.
Why ADR on its own is just as weak
ADR is room revenue divided by rooms sold — the average of what actually sold. Notice what is missing from the denominator: the rooms that did not sell.
So ADR rises when you turn business away. Sell three suites at $400 and refuse everything else and your ADR is $400. Your hotel made $1,200.
This is the trap in "we protected rate." Protecting rate on a night with real demand is discipline. Protecting rate on a night with no demand is just an empty hotel with a flattering average.
RevPAR is the referee
RevPAR — revenue per available room — is room revenue divided by rooms available. It equals ADR × occupancy, so both failure modes above are caught automatically.
Discount to fill and ADR drops faster than occupancy rises, so RevPAR falls. Hold rate and refuse volume and occupancy drops faster than ADR climbs, so RevPAR falls. The only way up is to sell more room revenue per room you own.
That is why it became the default. It is also why it is the number your owner, your lender and your benchmarking report all speak.
Where RevPAR stops being enough
RevPAR counts room revenue only. It does not know whether the booking arrived direct or through an OTA — an online travel agency such as Booking.com or Expedia. It also does not know what the room cost you to clean.
Add both layers to the same three scenarios. Assume a cost per occupied room (CPOR) of $45, which sits inside the midscale range Lighthouse publishes as an estimate ($40–65). Assume 18% OTA commission, matching the average acquisition cost measured by AH&LA, STR and Kalibri Labs. And assume the obvious thing — that the last rooms in a full house are the ones you had to buy through an OTA.
| Room revenue | OTA rooms | Commission | NRevPAR | Variable cost | Contribution per available room | |
|---|---|---|---|---|---|---|
| A — 100 @ $150 | $15,000 | 40 | $1,080 | $139.20 | $4,500 | $94.20 |
| B — 50 @ $200 | $10,000 | 20 | $720 | $92.80 | $2,250 | $70.30 |
| C — 80 @ $190 | $15,200 | 20 | $684 | $145.16 | $3,600 | $109.16 |
A and C were two dollars apart on RevPAR. They are $14.96 per available room apart on contribution — about $1,500 on the night, and A is the loser.
NRevPAR is net revenue per available room: room revenue minus distribution cost, divided by rooms available. GOPPAR is gross operating profit per available room, which takes it further by removing operating cost as well. The table above is a simplified stand-in for GOPPAR, not the full USALI calculation.
What the evidence says
Two findings matter here.
First, on adoption. A 2018 study by HSMAI APAC with the Singapore Institute of Technology measured which metrics hotels actually use: RevPAR 77.4%, RGI 48.5%, GOPPAR 20.4%, TRevPAR 13.7%. GOPPAR is the better decision metric and is used by one hotel in five, because no universal external benchmark exists for it. RevPAR persists because everyone calculates it identically.
Second, on the temptation to buy occupancy. Cornell researchers analyzed 67,008 hotel observations from 2001 to 2007 (Enz, Canina & Lomanno, Competitive Hotel Pricing in Uncertain Times, Cornell Hospitality Report Vol. 9 No. 10, 2009). Hotels priced 20–30% below their comp set ran occupancy 15.2% higher and RevPAR 12.2% lower. Hotels priced 5–10% above ran occupancy 0.5% lower and RevPAR 6.7% higher.
The authors state plainly that this is correlation, not causation. It is still the largest dataset anyone has published on the question, and it points one way.
One more number for scale. HotStats measured Americas flow-through at 20% for the year to August 2025, reported via HOTELS Magazine — for every extra dollar of revenue, about twenty cents reached profit. Revenue you buy expensively barely survives the trip to the bottom line.
So which one do you chase?
Chase RevPAR as your headline. Read occupancy and ADR underneath it to know how the RevPAR was made. Then check NRevPAR or contribution whenever the RevPAR answer is close, because a close RevPAR race is almost always decided by acquisition cost.
Put practically: RevPAR is what you report, contribution is what you decide on.
The mistakes that show up in these reports
Reporting occupancy first. Whatever leads the report becomes the target. Lead with RevPAR.
Comparing ADR across hotels with different fee structures. Under the 12th edition of USALI, compliance date 1 January 2026, resort and destination fees are Miscellaneous Income. They are not room revenue and they do not raise ADR.
Counting complimentary rooms as sold. STR excludes comps from both rooms sold and ADR. Include them and neither number will reconcile with your benchmark report.
Treating a full house as a win by default. Sold out early usually means you were the cheapest room in the market before demand had finished arriving.
Ignoring channel mix entirely. Two identical RevPAR nights can differ by thousands over a month, purely on where the bookings came from.
The bottom line
RevPAR is the right headline metric because it is the only one of the three that cannot be faked by neglecting the other. Use occupancy and ADR as the diagnostic pair that explains it. When two options land close on RevPAR — and they usually do — go one layer down to NRevPAR or contribution. That is where commission and cleaning cost finally get counted.
When you want ADR, RevPAR and occupancy sitting next to the market's, our analytics and reports module puts them in one view with MPI, ARI and RGI already calculated. It exports to HTML, PDF or Excel for the owner conversation.