On August 8, 2023, the Intecracy Group consortium held an online event in the Intecracy Expert Webinar format. The announced topic was “Master Data Management and Reference Data Management,” and the speaker was Serhii Balashuk, a leading solution architect. His presentation framed a system for managing enterprise reference and master data as a practical way to remove duplicate records, align business directories across applications and create a Single Source of Truth for operational decisions.
For organizations working with enterprise document management, electronic archives, system integration and Data Governance, the subject is highly practical. A document may be registered correctly from a formal standpoint, yet still carry process risk if the counterparty, product, address or identifier in a connected application differs from the corresponding reference data elsewhere. Such inconsistencies lead to reporting errors, manual checks and delays in decision-making.
Data quality behind controlled enterprise processes
Serhii Balashuk described how large organizations accumulate information across finance, logistics, sales, procurement and other functional areas. When each application maintains its own directories, the same business object can appear under several names, codes or sets of details. For an enterprise, this is not only an IT inconvenience. It affects reporting reliability, approval workflows, analytics and the cost of maintaining integrations.
“Creating a single source of truth is not a one-time technical project, but an ongoing management process. The main compromise that architects face when implementing MDM is choosing between rigid centralization, which slows down the entry of new records, and analytical consolidation, which allows systems to operate autonomously but requires complex algorithms for matching and merging data post factum. We must find a balance that corresponds to the dynamics of a specific business”, — noted Serhii Balashuk.
Deduplication and the trusted enterprise record
A separate part of the webinar addressed the mechanisms used to clean reference data. The process starts with standardization: addresses, phone numbers, company names, tax codes and other attributes have to be brought into a common format. After that, heuristic methods and fuzzy matching help identify records that are similar in meaning but not identical in their formal representation.
Once potential duplicates are detected, the system applies survivorship rules. These rules define which source has priority for a particular attribute. For example, a counterparty address may come from the CRM system, while payment details may be taken from ERP. This produces a Golden Record: an agreed record that contains the most current and verified information available across the enterprise landscape.
MDM architecture: registry, consolidation and centralization
During the presentation, three core MDM architecture models were reviewed. A registry keeps pointers and cross-reference identifiers in a central hub while source systems continue to store their data locally. Consolidation regularly collects information from local systems into a central hub for analytics and reporting, while operational input remains decentralized. The centralized, or transactional, model moves the creation and modification of master data into the MDM system, which then distributes updates to other applications.
The choice of architecture, as follows from the webinar, depends on the technical maturity of the IT landscape and on the organization’s readiness to change business processes. From an integration perspective, the important point is that future applications can connect to an already cleaned and structured data source instead of introducing yet another isolated directory.
Data Governance and responsibility for reference data
Technical tooling alone does not replace management rules. Serhii Balashuk emphasized the role of Data Governance: regulations, roles and responsibilities for creating, verifying and approving reference records. Data Stewards act as a bridge between IT and business users, especially when automatic deduplication cannot clearly resolve a conflict or decide whether two records are truly duplicates.
For enterprises developing document workflows, archives and integration environments, this approach provides a controlled data foundation for processes and reporting. Further context is available in the Intecracy Group publication about Serhii Balashuk’s webinar