Author note: Every week, I scan the tourism and hospitality industry for reports, updates and changes that could affect how hotels are discovered, marketed and booked.
What changed in hospitality, tourism, search and AI this week
5–12 September 2026
Vietnam wants to attract higher-value travellers, not simply more travellers.
Chinese travellers are increasingly using AI to decide which destinations even make the shortlist.
Meta’s newest AI agent can already move from planning into travel booking.
Airbnb is changing the economics of how its own inventory is priced and presented.
This week is a good reminder that hospitality marketing isn’t just about generating demand.
What happens to that demand from discovery all the way to transaction is more important.
Here are the signals I think are worth keeping.
Market & demand
🔥 Vietnam has changed the tourism scoreboard
Vietnam’s new tourism action programme sets some enormous targets for 2030:
- 45–50 million international visitors
- US$80–90 billion in tourism revenue
- 10–12% direct contribution to GDP
- Top three in Southeast Asia for tourism scale and growth
More important than the headline targets is the direction behind them.
Vietnam says tourism development should move toward greater quality, efficiency, sustainability and added value, with more attention on higher-spending and longer-staying travellers.
The signal: Vietnam doesn’t only want more demand. It wants better-value demand.
That should change how hotels evaluate their source markets.
Not simply: How many travellers are coming?
But: How long do they stay? How much do they spend? When do they travel? Which channels do they book through? And how much of that value reaches the hotel directly?
More tourism revenue for the destination does not automatically mean more direct revenue for an individual property.
Source: Government of Vietnam — Vietnam targets 45–50 million foreign arrivals by 2030
🔁 Vietnam remains one of Asia’s strongest destinations for repeat visitors
Agoda’s latest ranking places Vietnam among the top three Asian destinations for repeat international visitors.
Da Nang climbed two places to sixth among Asian cities, while the strongest Vietnamese destinations for repeat visitors included Da Nang, HCMC, Nha Trang, Hanoi and Phu Quoc. The ranking is based specifically on Agoda booking behaviour rather than all visitor arrivals.
I think the commercial implication is more interesting than the ranking.
The signal: A destination can earn loyalty without an individual hotel earning it.
Someone returning to Da Nang doesn’t necessarily return to the same hotel.
If both the first and second bookings happen through an OTA, the traveller may be loyal to Da Nang while the hotel itself owns very little of the relationship.
For destinations with strong repeat demand, CRM, remarketing, guest data, brand search and direct rebooking become particularly important.
Source: Da Nang Online — Vietnam among Asia’s top three destinations for repeat visitors
🇺🇸 US hotel RevPAR jumped 16.1%
For 30 August–5 September, US hotels recorded:
- Occupancy: 63.0%, +9.4% YoY
- ADR: US$159.19, +6.1%
- RevPAR: US$100.31, +16.1%
But CoStar specifically noted that the Labor Day calendar shift boosted the year-on-year comparison.
And the national number was far from universal.
St. Louis, for example, recorded an 11.8% decline in RevPAR.
The signal: A strong national benchmark doesn’t automatically mean strong demand for your property.
Before seeing +16.1% RevPAR and deciding rates can rise or acquisition budgets should expand, hotels still need to compare: equivalent holiday periods, local performance, booking pace and property-level data.
Source: Hospitality Trends — CoStar US weekly hotel performance
Direct booking & distribution
🛒 Hotel websites may have a conversion problem hiding behind a traffic problem
GuestCentric analysed data from 310 independent hotels during the first half of 2026.
It says that among guests who engage with its AI Website Sales Assistant, more than 56% show high booking intent, voluntarily sharing things like travel dates, group size and room preferences.
Meanwhile, GuestCentric describes roughly 2% as a typical hotel website conversion rate.
Those two percentages describe different groups, so I wouldn’t say that 98% of high-intent travellers abandon.
But the underlying point is still important.
Guests can arrive with strong intent and still leave because of things like:
- Unclear cancellation policies
- Hidden or late-surprise fees
- Confusing room descriptions
- Buried answers
- Rigid checkout forms
GuestCentric says some hotels are breaking beyond the traditional 2% ceiling toward consistent 10% conversion, while relevant, context-aware add-ons have increased average booking values by more than 38% in its research. These are vendor-reported results, rather than an independent industry benchmark.
The signal: Sometimes the direct-booking problem isn’t traffic. It’s what happens after the traveller arrives.
Hotels can spend more on: SEO → PPC → metasearch → social and still lose the guest right at the finish line.
Improving conversion makes every dollar spent generating demand more valuable.
This is an important part of Demand to Direct: Getting someone onto the hotel website isn’t the outcome. Getting the booking is.
Source: GuestCentric — Why do 98% abandon your hotel website?
🏠 Airbnb’s move to a 15.5% host-paid fee reaches another major deadline
Airbnb is continuing its move away from its traditional split-fee structure — where hosts typically paid around 3% and guests paid an additional service fee — toward a single 15.5% service fee paid by the host.
Airbnb’s official host communication says the deadline is:
- 15 September — outside the EEA
- 13 October — within the EEA, with exceptions in some markets.
Under the new model, guests see the price set by the host rather than seeing a separate Airbnb guest service charge added later.
For example, Airbnb illustrates the change like this:
Previously: Host sets £100 → guest sees about £115 → host earns £97
Under the single-fee model: Host sets £115 → guest sees £115 → host earns about £97.
The signal: Airbnb isn’t removing the distribution cost. It’s moving where that cost appears.
For hosts and property managers, that means margin and pricing strategy need to be revisited.
For hotels, I think the more interesting implication is price presentation.
Airbnb’s pricing experience becomes simpler because the guest no longer sees a separate platform service fee appearing at checkout.
So hotels cannot assume:
“Direct looks cheaper because Airbnb adds fees.”
The direct-booking proposition needs to be stronger than that:
- Transparent total pricing
- Better flexibility
- Direct-only benefits
- Better room/product information
- Loyalty
- Easier service
The battle isn’t just about headline rate.
It’s about total value and trust at checkout.
Sources: Airbnb — Simplifying service fees · Airbnb Official Community — September and October deadlines
🇪🇺 Google has changed how hotels and OTAs appear in European Search
I’ll keep this one simple.
For certain searches in the European Economic Area, Google now separates results into:
Aggregator results – OTAs, metasearch and comparison services.
Supplier results – direct providers such as individual hotels and hotel brands.
Google has also removed the dedicated Vacation Rentals unit from Search in the EEA, although feeds can still power other surfaces such as Maps and Google Hotels.
The signal: A platform redesign can change direct visibility without the hotel changing anything.
So if European organic traffic or direct bookings suddenly move, don’t automatically attribute the change to your hotel’s marketing alone.
Sources: Google — Supplier unit in Search · Google — Aggregator unit in Search
Hotels & resorts news
🏔️ InterContinental is bringing branded residences to Sapa
IHG and Alphanam announced The Residences at InterContinental Sapa Resort on 8 September.
The project will bring 79 branded villas to the Alphora Muong Hoa International Resort Complex and marks InterContinental’s branded-residence debut in northwest Vietnam.
The individual development is interesting.
The destination signal is more interesting.
The signal: International luxury hospitality investment is moving deeper into Vietnam’s secondary destinations.
Sapa isn’t Hanoi, HCMC, Da Nang or Phu Quoc.
International brands putting luxury inventory there suggests growing confidence in the destination’s ability to support a more global, premium traveller.
For independent properties, this raises the importance of owning the destination story.
A local hotel should theoretically know Sapa better than an international chain.
The challenge is making that knowledge equally visible across search, AI, social and its own website.
Source: IHG — InterContinental branded residences arrive in Sapa
AI Search & discovery
🇨🇳 44% of Chinese travellers use AI for planning — and AI is shaping the shortlist
Create Consulting China tested 240 travel queries across five major Chinese AI platforms.
The study, reported by Skift, found that AI’s biggest influence currently happens before the booking.
It helps determine which destinations even enter the traveller’s consideration set.
And 44% of Chinese travellers now use AI for travel planning, while OTAs still dominate the later comparison and booking stage.
The systems tended to favour destinations with: easy visa access, shorter flights, strong safety perception, more Chinese-language information online
Broad prompts frequently favoured already-established destinations such as Singapore, Thailand and Japan.
More specific queries produced very different shortlists.
The signal: AI doesn’t only influence which option gets chosen. It can influence which options get considered at all.
And there’s a feedback loop: Popular destination → more information → easier for AI to understand → more recommendations → more visibility.
For hotels and destinations targeting China, Chinese-language information coverage may increasingly become part of market access, not simply localisation.
Source: Skift — The race for Chinese travel demand is now a fight to get on the AI shortlist
🤖 Meta’s newest AI agent can already book travel
Meta launched Muse on 8 September.
Unlike a chatbot that simply gives an answer, Muse can browse the web and take actions on behalf of the user — including booking travel.
For flights, Meta has integrated with Duffel, allowing Muse users to search live inventory, compare options, make bookings and manage trips.
Accommodation isn’t part of that Duffel integration yet, although Duffel says it plans to support it in the future.
The signal: The AI travel conversation is moving from recommendation to transaction.
Sources: Meta — Introducing Muse · Duffel — Muse travel integration
⚠️ AI-generated fake hotel videos are becoming a real brand problem
Skift investigated AI-generated TikTok videos promoting Ambiente Sedona, a luxury resort in Arizona.
Videos showed suites and even a “peaceful pond” that the property’s owner says do not exist.
Some creators also appeared to be AI-generated, while commission tags did not always match the actual booking reality.
One fabricated video had already received more than 315,000 views.
TikTok Go makes this more important because travel content can increasingly sit directly beside booking and commission mechanisms.
The signal: Hotels are losing some control over their own brand facts.
Sources: Skift — Fake AI Hotel Videos Are Coming for Travel · Skift — Your Hotel Is Hot on TikTok Go. But AI Got All The Facts Wrong.
So what?
Look at the whole journey this week.
The hospitality marketing funnel is starting to look less like: Visibility → website → booking
and more like:
Demand → discovery → shortlist → trust → comparison → transaction → direct relationship
The hotel doesn’t need to own every step.
It probably can’t.
But there’s a huge difference between participating in that journey and allowing platforms to own almost all of it.
And perhaps that’s the simplest way to explain what I’m trying to explore with this newsletter: Demand already exists.
The interesting question is how much of it a hotel can turn into direct business.
That’s Demand to Direct #03.
I’ll be doing this weekly: the tourism, hospitality, search and AI updates I think are actually worth paying attention to with the source, the context and one thought on why it matters.
See you next week.