Richemont's serial number problem and how product-level data closed the grey market gap

In 2019, Richemont made an acquisition that most analysts read as a digital commerce play: the buyout of the remaining stake in Yoox Net-a-Porter. What received less attention was a parallel investment in Richemont's own authentication and track-and-trace infrastructure, specifically for its hard luxury brands, Cartier, IWC, Panerai, and Vacheron Constantin among them. The driver was not counterfeiting in the conventional sense. It was authorised retailers in lower-duty markets, particularly in Hong Kong and Dubai, selling allocated pieces into grey-market channels that then resurfaced in mainland China, Japan, and South Korea at prices that undercut the official retail network. The pieces were genuine. The problem was entirely one of distribution control, not IP violation.

This is a meaningful distinction for a CDO. Counterfeiting is, at its core,a legal battle fought through customs seizures, trademark enforcement, and cross-border litigation. Grey-market leakage from authorised partners is different: no law is being broken by the retailer, and the brand cannot simply litigate its way out. The data problem has to be solved operationally, at the level of individual product units moving through channels.

What Richemont actually built

The architecture Richemont assembled across 2020 to 2023 combined three data layers that had historically sat in separate systems.

The first was unit-level serialisation. Richemont's watch brands already had movement serial numbers, but those numbers were not consistently tied to point-of-sale records in a queryable format. The group invested in harmonising those identifiers across its retail management systems, creating a single record per watch that tracked manufacture date, initial destination market, authorised retailer assignment, and sale date. For pieces sold through wholesale partners rather than directly, sell-in and sell-out data were reconciled: the moment a piece moved from a retailer's inventory to a named buyer, that event was logged.Reconciling sell-in and sell-out signals is structurally harder in watch retail than in fashion because transaction volumes are low, average selling prices are high, and wholesale partners have historically resisted sharing granular sell-out data, treating it as a negotiating asset.

The second layer was secondary-market monitoring. Richemont built, partly through Watchfinder (which it acquired in 2018), a data feed that tracked asking prices and transaction records on the major grey-market and pre-owned platforms: Chrono24, WatchBox, and various regional resale marketplaces in Japan and China. By matching secondary-market serial numbers against the primary-sale database, the team could identify not just that a piece had leaked, but which authorised retailer had sold it and into which geography it had moved before resurfacing. This is the analytic that changes the conversation with wholesale partners from accusation to evidence.

The third layer was anomaly detection on allocation patterns. Richemont's maisons manage demand through waiting lists and controlled release of hero references. Where a retailer was receiving its normal allocation but clearing inventory faster than local foot traffic would explain, that delta became a flag. The logic is simple: if an authorised dealer in Geneva sells twelve Cartier Santos pieces in a quarter and its local clientele file contains eight verified Swiss buyers, the remaining four require explanation.

The AML and KYC infrastructure that Richemont already maintained for high-value transactions (required under Swiss and EU anti-money-laundering rules for watches above certain price thresholds) provided a complementary data source. Buyer identity records, where they existed, could be cross-referenced against secondary-market seller profiles on platforms that required registration.

Results: what is known and what is uncertain

Richemont has not published grey-market leakage rates as a standalone metric, so any figure here would be invented. What is documented in the group's public communications and analyst briefings is the following:

By the 2023 financial year, Richemont reported that its directly operated retail share of revenue had risen significantly relative to wholesale, reaching roughly 68 percent of group sales from hard luxury. That shift is partly strategic (the group has been buying back concessions from multibrand retailers), but it also reflects that the data infrastructure made it possible to identify and in some cases terminate wholesale relationships that were demonstrably feeding grey markets.

Watchfinder's integration into the intelligence loop is harder to quantify. The platform processes tens of thousands of pre-owned watch transactions annually, and the serial-number matching capability it provides is a genuine operational asset, though Richemont has been careful not to overstate its coverage, given that a significant portion of grey-market transactions move through non-digital channels in mainland China.

What transfers to your context

Three things from this case are portable across luxury groups, with caveats.

Unit-level serialisation only works if your wholesale contracts require sell-out reporting. Without that contractual lever, your primary-sale database is incomplete and the matching exercise produces false negatives. If your current partner agreements do not include sell-out data obligations, the next contract renewal cycle is the moment to change that. Retailers will resist, but the argument is increasingly supported by the group's ability to demonstrate mutual benefit: a brand that can prove allocation is going to genuine end-buyers is a brand that can justify premium placement and co-investment.

Secondary-market monitoring requires a deliberate decision about what you do with the intelligence. Knowing that a specific authorised retailer in Singapore is a consistent source of grey-market leakage is actionable, but the action is a commercial and legal decision, not a data decision. The CDO's job is to make the evidence unambiguous enough that the commercial team cannot ignore it. That means documented methodology, reproducible matching logic, and a clear confidence interval on each flag.

The AML angle is underused in most luxury data strategies. The transaction records generated by high-value sale compliance processes contain buyer identity and payment method data that, treated correctly under GDPR and equivalent regimes, can materially strengthen grey-market attributionattributionA framework for assigning credit to the touchpoints that contributed to a conversion, so you can measure which channels and interactions actually drive results.View full definition →. Most luxury groups have legal teams that manage AML data in isolation from the commercial analytics function. Bridging that gap requires a governance conversation, not a technical one.

Where context differs: this approach works for hard luxury goods with permanent serial numbers. Soft luxury, including leather goods and ready-to-wear, requires a different authentication layer, typically NFC chips or cryptographic tags embedded at manufacture, because the product itself does not carry a persistent unique identifier by default. RFID-based track-and-trace in leather goods is a separate problem with its own implementation economics.

The core lesson from Richemont is that grey-market leakage is a data qualitydata qualityThe degree to which data is fit for purpose: accurate, complete, consistent, timely, valid and unique. Poor quality data undermines analytics, reporting and AI.View full definition → problem before it is anything else. If you cannot match a unit from manufacture to end-buyer with confidence, you cannot distinguish a leaking partner from a well-performing one. Building that match is the first investment, and it pays before you run a single anomaly detection model.

The full course on this sector:Data in Luxury.