When an organization has the same customer listed five times in its CRM with slightly different names and addresses, and three different employee records for the same person in its HR system, and the same vendor appearing in four purchasing systems with different payment terms in each — it has a master data problem. And master data problems are not data problems in the abstract. They are operational problems: wrong deliveries, duplicate payments, incorrect reporting, failed integrations, and compliance failures that trace back to the absence of a single, authoritative record for each business entity.
What Master Data Actually Is
Master data is the core reference data that the business runs on — customers, products, vendors, employees, organizational units, and the other entities that appear across multiple systems and processes. It is distinct from transactional data, which records business events and references master data entities. The customer master record is master data. The sales order that references a customer is transactional. When the customer master record is inconsistent, every transactional record that references it is potentially incorrect.
The Four MDM Problems to Solve
Master data management addresses four problems that, if left unaddressed, compound each other. Duplicate records — multiple records representing the same real-world entity — need to be identified, merged, and prevented from reoccurring. Incomplete records — master records missing data required for downstream processes — need to be enriched and validated before they enter production systems. Inconsistent records — master data maintained differently in different systems with no mechanism to synchronize them — need a defined single source of truth and a propagation mechanism. And ungoverned creation — the absence of process controls that prevent new records from being created without meeting quality standards — needs a governed workflow that enforces completeness and accuracy at point of creation.
Starting Smaller Than You Think You Need To
Enterprise MDM programs that attempt to solve all four problems simultaneously across all entity types in all systems typically fail under their own weight. The MDM initiative that produces durable results starts with one entity type — usually the most problematic, or the one with the clearest business impact — and implements governance, cleansing, and process controls for that entity before expanding to others. The discipline of starting small and proving value before expanding is what distinguishes successful MDM programs from expensive projects that were eventually abandoned.