Master data management for customer, product

Problem class

The same customer exists as five different records across CRM, ERP, e-commerce platform, loyalty system, and support tool — with different names, email formats, and IDs. Post-merger acquisitions bring duplicate product catalogs with overlapping SKUs. Finance closes cannot reconcile because customer IDs don't match across systems. Marketing campaigns hit the same person multiple times under different identities. Customer service sees an incomplete interaction history. KYC and compliance checks cannot confirm whether a customer is the same person as a known bad actor. Gartner reports 75% of MDM programs fail to meet business objectives.

Mechanism

MDM establishes golden records by: (1) ingesting entity data from all source systems into a staging layer, (2) applying probabilistic + deterministic matching rules to resolve duplicates (same person, same product), (3) applying survivorship rules to determine which attribute value "wins" in the golden record (most recent, most authoritative source), (4) publishing the golden record back to consuming systems via APIs or data platform connectors, and (5) operating a stewardship UI for manual resolution of ambiguous matches. AI-assisted matching (Informatica CLAIRE, Tamr's ML matching) dramatically reduces the manual matching workload.

Required inputs

• Inventory of all systems containing the target master entity (customer, product, supplier)
• Unique identifier mapping across systems
• Matching rules (deterministic: same email = same person; probabilistic: fuzzy name + address)
• Survivorship rules (which source wins for each attribute)
• Stewardship process and staffing for manual review queue
• API or connector layer for golden record distribution

Produced outputs

• Golden record per entity with confidence score and source attribution
• Deduplicated entity count and match statistics
• Customer 360 or Product 360 view for downstream consumption
• Reduced compliance risk (KYC, GDPR right-to-erasure, AML)
• Clean foundation for CRM, personalization, and analytics

Industries where this is standard

• Retail/commercial banking and insurance (Citizens Bank, Empire Life) where KYC, regulatory compliance, and customer 360 views are mandatory
• Healthcare and life sciences (IQVIA) with HCP/HCO mastering and HIPAA compliance
• Hospitality and travel (JetBlue, Holiday Inn Club Vacations) with guest profile unification and loyalty programs
• Large retailers with omnichannel customer data and product information
• Post-M&A manufacturing (Imerys, Yamaha) requiring data consolidation across acquired entities

Counterexamples

• Single-system businesses: Organizations with one CRM and one ERP with no M&A history don't need MDM — there's nothing to deduplicate.
• Multi-domain, enterprise-wide rollout from day one: Attempting to master customers, products, and suppliers simultaneously is an anti-pattern. Start with one domain, one business unit.
• Treating MDM as an IT project: Without C-level authority and business ownership, cross-departmental resistance consistently stalls progress. 90% of businesses fail on their first MDM attempt (Gartner).

Representative implementations

• Citizens Bank (Informatica Cloud MDM) reduced data onboarding time by 85% using AI-powered automation, moved from batch processing (4–15 hours) to near real-time, and reduced customer profile update latency from up to 3 days to instant.
• JetBlue maintains nearly 100 million unified customer profiles, processing 1M+ events daily in real time, with 99% of customer interactions tied to unified profiles — reducing service friction and eliminating outages.
• Rodobens (Brazilian financial services, Informatica MDM) achieved 182% of GMV target in the first six months of launching CRM campaigns on unified data, projecting an additional ~$40M in annual GMV, with 50%+ reduction in IT hours for monthly data maintenance.
• Tamr (Forrester TEI composite) demonstrated 643% ROI over three years (~$9M total benefits vs. $1.18M costs), with 50%+ shorter customer onboarding at financial services organizations.

Common tooling categories

MDM platform (Informatica MDM / Tamr / Reltio / Semarchy / Stibo) + matching engine (probabilistic ML + deterministic rules) + survivorship engine + stewardship UI + golden record API / CDC distribution layer + data quality scoring.