Douyin's 2m views, under 20% direct bookings: where hotels leak revenue
A hotel's Douyin campaign drew 2m views but few direct bookings. The gap sits between the click and the reservation, not in ad spend.

Douyin's 2m views, under 20% direct bookings: where hotels leak revenue
A hotel's Douyin campaign drew 2m views but converted fewer than one in five into direct bookings. The money did not leak in the ad buy; it leaked between the click and the reservation.
Our view is simple: many hotels do not lack guests. They lack a path that turns a click into a booking.
Exposure does not buy orders; the problem is not the ad buy
Look at the detail in that data set. The hotel launched its first Douyin local-push campaign in November 2025, using a "generic landing page plus manual rate adjustment" model. The landing page was a static H5, with the price fixed at 10% off the rack rate and no link to live inventory or dynamic room rates. After clicking, users were redirected to a third-party booking engine, and the first-screen bounce rate was 68.4%. Business travellers, family guests and Gen Z visitors were all pushed the same room-type combination.
Put those three together and they point to one fault: the ad side is spending money to buy attention, while the receiving side is a "dead" page. The price is dead, the inventory is dead, and the audience is undifferentiated.
According to the article's account, after the hotel connected a revenue management system, the direct-booking conversion rate rose from 18.5% to 55.7%. Those numbers belong to someone else and need not be copied, but they show one thing: the traffic did not increase; what changed was what happened after each click.
For accommodation operators, this is a trade-off: keep adding budget to buy exposure, or first fix the stretch after the click. Pour budget into a leaking bucket and it leaks as much as you pour in.
Who is affected, and who does not feel it yet
The most directly affected are mid-to-high-end independent hotels and boutique hotels. These hotels have a fair number of rooms and high service requirements, but they lack the central reservation and membership systems of a chain group. Advertising, reception, room inventory and pricing are often managed by several different teams, coordinated through WeChat and spreadsheets.
According to a 36Kr article on hotel membership and OTAs, a decade ago the hotel industry relied heavily on OTAs for guest acquisition. In a stock market, the "talent grab" of hotel groups upgrading their membership systems may be grabbing the OTA platforms' customers. The article mentions a scenario: a business traveller books on an OTA and then calls the hotel to ask about invoicing. The hotel says that booking through its own membership programme not only offers a room-rate discount but also makes invoicing easier, and the guest ends up bypassing the OTA.
This shows guests are not unwilling to book direct; they need a smooth reason and a smooth path. If the path is blocked, the reason is wasted.
Where the receiving stage gets stuck, action by action
Break "click to order" into steps: the guest sees the price, confirms whether a room is available, decides which room type to book, completes the booking, and checks in.
The problem with a static landing page is that the price and room inventory are already out of date the moment the guest clicks in. Room rates should move with supply and demand, inventory should be precise down to how many rooms are left that day, and room types should be recommended by guest type. Manual rate adjustment and one generic page cannot do these three things.
SHEYU AIHotel's corresponding capabilities along this chain are: intelligent rate management using a BAR base rate, negotiated rates, seasonal rules and a real-time simulator, with data-driven revenue management; visualised room and inventory status, with a booking-to-checkout loop covering the full status tracking of creating, modifying and cancelling reservations, check-in identity verification, room-folio posting and settlement, and deposit management. The guest CRM profile can match returning guests through hashed ID numbers and automatically identify VIP levels and preference profiles.
In other words, the price and inventory a guest sees after clicking are backed by a system, not a graphic made yesterday.
Then there is the service stage. Phone calls transferred to a human no longer require jotting down the room number by hand. WeChat, tablet and official account share one session, and when a guest switches from the in-room tablet to a phone, the AI remembers the context. Back-office staff handle the transferred work order rather than understanding the request from scratch. A multilingual knowledge base is part of this.
What it cannot solve
To be clear: this system will not decide for you whether the ad budget should go to Douyin or to OTAs, nor will it conjure traffic out of thin air. It manages the stretch after traffic arrives: whether inventory is accurate, whether prices move, whether room types are recommended correctly, and whether someone picks up what guests ask. Front-end acquisition strategy remains the operator's own judgment.
It is also aimed at high-end hotels with hundreds of rooms. Homestays or small independent properties with very few rooms may not need such a system.
Where small and medium businesses can start now
There is no need to rebuild the whole system at once. Start by asking yourself three questions:
- On the page guests click into, are the price and room inventory real-time?
- Do guests from different sources see the same set of room types?
- When a guest moves from the platform to WeChat and then to a phone call, does anyone still remember what was said before?
If you cannot answer one of these three, that is where the leak is. Fix that stretch first, then talk about whether to add budget.
Traffic is rented; the ability to receive it is your own.