Its own platforms, its own retailer,
and no room for a generic plan.
Expanding into an ecosystem they had not run in.
Cartier Travel Retail was expanding in Hainan and Macau, with three things to do at once: promote new boutiques, drive traffic to a WeChat loyalty mini-app, and build awareness across Chinese social platforms.
The brand had limited operational experience with the Tencent ecosystem and needed a partner to plan, execute and report year-round campaigns across the Watches, Fragrance and Jewellery lines.
Year-round, per line, through the retailer.
We ran year-round media across the Tencent ecosystem, WeChat advertising and OTAs, targeting Chinese travellers in Hainan and Macau. Audience targeting was customised per product line and kept aligned to Cartier's wider brand messaging rather than treated as three separate accounts.
We worked closely with the duty-free retailer CDFG, monitored performance continuously, and supported the WeChat mini-program as the conversion destination. Year-round matters here: a Chinese travel retail calendar is not a series of bursts.
The Chinese New Year activation in February 2021 ran as an explicit test: three audience definitions bought side by side, across search, display, social and SMS, to find out which one a travel retail brand should actually buy. It answered two questions worth carrying into every campaign since.
On the campaigns that followed we ran independent ad verification alongside the vendors' own reporting, rather than accepting the numbers the people selling the media chose to report. In a market where the buyer usually has no access to the platform and no view of how the targeting was set, that is the only way to know what was actually delivered.
A market with its own architecture.
The first year-round Tencent ecosystem campaign we ran for a luxury jeweller expanding into Hainan, a market that punishes a generic media plan.
The test settled where the traveller data has to come from. A Chinese vendor's own confirmed-traveller segment underperformed the pre-trip intent audience, and the report is explicit about why it could not be trusted: we had no access to the buying platform and no view of how that segment was actually built. Where the confirmed-traveller signal is a real booking, from an airline data partner, it is the strongest placement in the plan, which is exactly what it was for Nestlé at the World Cup and for Royal Salute at Taoyuan. The lesson is not to buy intent instead of confirmation, it is to know whose data is behind the label before paying for it.
The second finding: in Hainan almost every click came from generic search rather than the brand's own name, which means the campaign was reaching people researching where to shop at the destination rather than people already looking for the maison.
Independent verification earned its place immediately. One vendor's click reporting ran as much as a third above what the third-party measurement recorded, and drifted day by day, while another stayed close and steady. Delivery was also landing outside the target region until it was caught and corrected. Neither would have been visible from the vendor dashboards alone.
The last lesson is structural, and it is the one most brands miss. The domestic China team and travel retail ran near-identical creative into the same island at the same time, so the audiences overlapped and buyers could not tell the two offers apart. In travel retail the budget and the audience pool are always the smaller of the two, which makes coordination with the domestic team part of the media plan rather than a courtesy.
This page carries no performance figures on purpose. The activations were run as test phases, their own reports are lists of what to fix rather than scoreboards, and the workbook documents nothing we could stand behind.