Knowing Every Customer: Real-Time MDM for a Global Luxury Retailer
How a global luxury brand replaced a legacy deduplication engine with Tamr’s AI-native MDM to deliver a single, trusted view of its customers.
A global luxury retailer operates flagship stores across major international markets, serving some of the world’s most sought-after fashion and lifestyle customers. The retailer manages 11M+ customer records across in-store, e-commerce, CRM, loyalty, and clienteling solutions.
Challenges
- Homegrown customer data deduplication engine was hard to maintain and failed to handle merges and splits of records
- Fragmented customer data created duplicate profiles across channels and touchpoints, resulting in attribution gaps for high-value customers
- Revenue metrics were distorted due to an inability to link purchases to the correct customer
- Stringent data privacy requirements (including GDPR) created compliance risks
Outcomes
- Replaced legacy deduplication engine with Tamr’s real-time AI-native MDM without disrupting downstream systems
- Enabled search-before-create and search-after-update functionality to prevent duplicate customer profiles and support continuous remediation
- Automatically resolved two-thirds of duplicate customer records and provided an easy-to-use interface for human curation of edge cases
- Improved customer matching accuracy by 5–9% compared to the legacy system, helping ensure more purchases were attributed to the correct customer profiles
For this global luxury retailer, every customer relationship is a valuable asset. Customer advisors manage books of 1,000+ customers, yet a small elite segment drives most revenue. When purchases cannot be connected to the right customer, advisors lose visibility into spending patterns, and business leaders operate with incomplete information.
Our top spenders represent a small fraction of our customer base, but they carry an outsized share of our revenue. When we can’t link their purchases to the right record, we’re not just losing data—we’re making the wrong decisions about our business.”
The retailer’s customer data infrastructure had reached its limits. A homegrown deduplication engine could no longer be maintained. When the same customer appeared across multiple touchpoints—different stores, sales channels, or geographies—the result was duplicate accounts, lost spend attribution, and an incomplete view of valuable clients.
The stakes came into sharp focus when the retailer’s customer advisor expected a significant purchase—a single transaction worth tens of thousands of dollars—but the purchase occurred online without the customer authenticating against their existing profile. Because the legacy system could not merge accounts, the transaction was attributed to a “new” customer. Since a relatively small percentage of customers generated a majority of the revenue, accurate identity resolution for elite clients was a strategic priority.
To address these challenges, the retailer turned to Tamr to modernize its customer identity resolution capabilities and replace the legacy engine.
At the same time, the retailer had to balance this transformation with stringent operational and compliance requirements. Tamr needed to replace the existing solution without disrupting any of the nine downstream systems. Given GDPR obligations and internal governance requirements, any data change had to be incremental, auditable, and approved. Before deployment, the retailer’s Data Protection Office conducted a detailed review of security controls, governance processes, and data handling practices. Tamr’s flexible architecture enabled the retailer to meet strict compliance requirements without slowing modernization efforts.
Initially, Tamr was deployed in a dry-run mode alongside the existing engine, enabling side-by-side validation of match quality. The results demonstrated 5 to 9% higher matching accuracy than the legacy system and helped build confidence among internal stakeholders. Once validated, Tamr replaced the legacy engine as the system of record, with downstream systems consuming the output to support a more accurate and trusted customer view.
We didn’t need to rip and replace everything at once. Tamr could come on as a system of record and let us transition while giving us the flexibility to operate within our compliance boundaries.”
Within one week of the go-live date, approximately 2,100 potential duplicate candidates were identified. Two-thirds were automatically resolved, while the remaining records—with lower match confidence—were surfaced in Tamr’s Curator Hub for human review, giving data stewards visibility and control over edge cases.
Now, to support continuous improvement, real-time search-before-create APIs prevent duplicate records from entering the system. And search-after-update APIs automatically re-evaluate records when new information becomes available or matching logic changes, enabling continuous remediation across millions of records without manual reprocessing.
“Two-thirds of flagged duplicate candidates were resolved automatically in the first week. That means our model is precise enough that the business trusts it to act without anyone needing to review it.”
With a trusted customer identity foundation in place, the retailer is now focused on expanding the business value of customer data. Future initiatives include greater automation of stewardship workflows with Curator Hub, deeper customer relationship intelligence, and more personalized engagement strategies for the clients who drive the greatest share of revenue.
See for yourself
Get a free, no-obligation, 30-minute demo of Tamr, and discover how our unique AI-native MDM solution can empower you to deliver data you can trust.