Marriott Beats on Bonvoy and International — But U.S. RevPAR Quietly Dipped

Marriott-branded printed earnings release fanned out on a hotel lobby banquette

Marriott's Q3 beat the Street on EPS and EBITDA, but a -0.4% U.S./Canada RevPAR print and a record 3,923-property pipeline tell a more honest story about where the lodging cycle actually sits in late 2025.

I read the Marriott print the way I read a banquet check at 2 a.m. — top line first, then the part the captain hopes you’ll skip. The top line on Tuesday morning was clean: $2.47 adjusted EPS against a $2.37 consensus, $6.49 billion in revenue, adjusted EBITDA up 10% to $1.349 billion. The part the captain hoped you’d skip was buried two paragraphs into the official release: U.S. and Canada RevPAR fell 0.4% in the third quarter.

That is the number that matters, and it is the number nobody on the morning shows wanted to lead with. So let me be the captain who points at it.

The beat is real. The mix is the story.

Marriott’s worldwide RevPAR ticked up 0.5%. International did the heavy lifting at +2.6%. Luxury — the part of the portfolio that doesn’t flinch when a GSA per diem gets clipped — ran +4%. The U.S./Canada softness isn’t a leisure problem and it isn’t a corporate-transient problem in the aggregate. CEO Tony Capuano told the call it was a government-travel pullback, and the Hotel Dive write-up flagged that the federal shutdown overhang and lower agency travel volumes hit Washington-adjacent and secondary-market full-service properties hardest.

That tracks with what I’ve been hearing from F&B directors at airport-cluster Courtyards and Marriotts in the Mid-Atlantic since August: contract catering for visiting federal teams went quiet, group rebookings slid into Q1, and banquet covers softened just enough to spook the labor model. None of that shows up in a 50-basis-point worldwide RevPAR print. It shows up in the U.S./Canada line.

What rescues the quarter — and what should rescue the multiple — isn’t the RevPAR walk. It’s the pipeline. Marriott added 17,900 net rooms in Q3 and now sits on a record development pipeline of 3,923 properties and more than 596,000 rooms. That is, functionally, a second Marriott waiting to open. The earnings call transcript had Capuano leaning into conversions, the Sonder integration, and a midscale and extended-stay push that doesn’t need U.S. RevPAR to cooperate to keep compounding fees.

The captain’s tell

Here is the part I want to mark as interpretation, not reporting: a fee-driven model like Marriott’s is allowed to have a soft same-store RevPAR quarter as long as the pipeline keeps widening. The market knows this. That’s why the stock didn’t get punished on a print where the domestic comp went negative. $0.8 billion of Q3 buybacks and a $1.349 billion EBITDA line gives management the cover to call the U.S. dip “transitory” without anyone laughing.

But the read I’d push back on — and this is where I differ from the sell-side notes that hit my inbox this morning — is the idea that the pipeline alone makes the U.S. softness ignorable. Pipeline is a 2027–2029 story. The 2026 numbers still have to clear a U.S. lodging market where government travel is structurally lower, group is fine but not heroic, and luxury is increasingly the only segment doing real RevPAR work. Bonvoy enrollment growth and international room adds will paper over a lot. They won’t paper over a second consecutive negative U.S./Canada quarter if that’s what Q4 prints.

What I’m watching at the property level

Three things, all of which sit closer to my beat than to the analyst beat:

First, the Sonder integration. Marriott confirmed the rooms are flowing into the system, but the operational handoff — F&B-light, tech-heavy, brand-adjacent — is the kind of thing that either compounds quietly or generates a Reddit thread. I haven’t seen the Reddit thread yet. That’s a good sign.

Second, the AI deployment surface inside the Bonvoy stack. I’ll have more on this in a forthcoming May piece — /blog/posts/inside-marriotts-ai-stack-a-case-study-in-how-the-largest-hotel-chain-is-actually-deployin — but the short version is that the same fee model that lets Marriott absorb a -0.4% U.S. RevPAR quarter is also the model that makes every basis point of personalization-driven attach-rate worth disproportionate margin. The pipeline pays the rent. The app pays the bonus.

Third, the labor line in the U.S. full-service portfolio. If government travel doesn’t come back in Q1 — and the federal calendar suggests it won’t snap back the way operators want — banquet and catering staffing models built in 2023 are going to get rewritten before the spring shoulder. That’s an F&B story, not a RevPAR story, and it won’t show up on a press release.

The headline number was $2.47. The honest number was -0.4%. The number that actually matters is 3,923.

— Naomi covers hotel F&B and operator tech for TableTransfers. Tips: [email protected].

Featured More

The Voice Agent Maturity Curve

mise

·

12 min read

The Four Margins of a Restaurant

mise

·

14 min read

The AI Premium in Hospitality M&A: Broker Story or Real Number?

the bottom line

·

9 min read

What the DoorDash/SevenRooms Deal Actually Buys

the bottom line

·

11 min read

Browse all 494 posts

Related posts

Toast Quietly Renamed Sous Chef. The Pilot Was the Point.

the pass

·

6 min read

Toast Quietly Renamed Sous Chef. The Pilot Was the Point.

Darden +4.2% comps and the boring Bahama Breeze ending

the pass

·

5 min read

Darden +4.2% comps and the boring Bahama Breeze ending

Bahama Breeze is closing. Darden's portfolio thesis just got tighter.

the pass

·

5 min read

Bahama Breeze is closing. Darden's portfolio thesis just got tighter.