Master Data Management Services

Master data is the slow-changing core information that describes the business entities transactions refer to. An order is transactional data; the customer who placed it and the product it contains are master data. Reference data, such as country codes or status lists, is a related but simpler category.

An engagement usually covers choosing the domains to start with, profiling current data, defining the golden record, building matching and merge logic, connecting source systems, and setting up stewardship so quality stays high after go-live.

• Customer or party: people and organizations, including relationships such as parent company and contact.
• Product or item: descriptions, hierarchies, units and attributes shared by sales, finance and supply.
• Supplier or vendor: legal entities, banking details and risk status, where duplicates can cause duplicate payments.
• Location and asset: sites, warehouses and equipment referenced by operations.

Start with one domain where bad data visibly costs the business, such as failed deliveries or double billing. A narrow first release proves the approach and earns the trust needed to extend it.

Building a golden record follows a repeatable sequence:

1. Standardize. Clean formats for names, addresses, phone numbers and product codes so comparison is fair.
2. Match. Decide which records describe the same entity, using exact rules (same tax number) and fuzzy rules (similar name and postcode). Machine-learning matching can help with messy data, but its borderline cases still need review.
3. Merge with survivorship rules. For each field, choose which source wins, for example billing system for legal name and CRM for contact email.
4. Publish. Send the golden record, with its cross-reference to every source ID, back to the systems that use it.
5. Review exceptions. Uncertain matches go to a human queue instead of being merged automatically.

Wrongly merging two different customers is usually worse than leaving two records unmerged, so tune the match thresholds toward caution and record every merge so it can be reversed.

Software merges records, but people decide what correct means. Name a business owner for each domain, and data stewards who resolve exceptions, approve changes to survivorship rules and answer questions from users. Without these roles, duplicates creep back within months.

Associations such as DAMA International (opens in a new tab) publish widely used guidance on data management roles and practices. For the broader operating model, see our data governance consulting service.

Master data has to reach the systems where work happens. Plan how the hub connects, with help from data engineering where pipelines are needed, to the CRM, ERP, e-commerce platform and warehouse, using APIs for live lookups and batch feeds for analytics. Conformed dimensions, a concept explained in the Kimball dimensional modeling techniques (opens in a new tab), are how the golden record becomes shared customer and product tables in reporting.

• Automated tests on uniqueness, completeness and valid values, using tools such as dbt tests (opens in a new tab) or Great Expectations (opens in a new tab).
• Quality metrics visible to owners, such as the share of customers with a verified address, tracked over time.
• Change history, so you can see who altered a record and why.

A single customer view concentrates personal data in one place, which raises both the value and the risk. Plan for access limits, retention rules and the ability to find and correct or delete everything held about one person. Confirm details with legal counsel.

USA: the California Consumer Privacy Act (opens in a new tab) gives residents rights to know, delete and correct, and a hub makes honoring them easier. UK: the ICO’s UK GDPR guidance (opens in a new tab) includes accuracy and storage limitation principles that master data work directly supports.

Australia: the OAIC’s APP guidelines (opens in a new tab) address data quality, security and access to personal information. Singapore: the PDPC’s overview of the Personal Data Protection Act (opens in a new tab) covers accuracy, protection and retention obligations. Kramate offers four to six hours of overlap with US Eastern time and a full working day with the UK, Australia and Singapore for stewardship workshops.

• Number of domains and source systems, and how inconsistent they are.
• Data volume and the difficulty of matching, for example records without shared identifiers.
• Architecture style: registry is lighter than centralized.
• Platform licensing versus a custom-built hub.
• Stewardship staffing, which is an ongoing business cost, not a project cost.
• Integration and change effort in every consuming system.

Fixed price works for a single domain with known systems, a dedicated team for a multi-domain program, and hourly work for assessments. Quotes for master data management services should state which of these they include. Always budget time for data cleanup before migration into the hub, since this is where plans most often slip.

1. Domain-first thinking: they start from a business problem, not a product demo.
2. Vendor neutrality: they can explain build versus buy and when a lightweight approach is enough.
3. A stewardship plan, not only technical design.
4. Match quality evidence: they describe how they measure false merges and missed matches.
5. Reversibility: every merge can be traced and undone.
6. Ownership: you receive the rules, mappings and documentation.

A dedicated project manager coordinates the work, with daily updates in Slack or Teams and a progress review closing each two-week round. We can lead delivery or embed engineers in your data team, under an NDA signed before detailed talks, with all intellectual property transferred to you. Commercial options are fixed-price, dedicated-team and hourly. See how our delivery approach works.

• DAMA International (opens in a new tab)DAMA International
• Dimensional Modeling Techniques (opens in a new tab)Kimball Group
• dbt Documentation (opens in a new tab)dbt Labs
• Great Expectations Documentation (opens in a new tab)Great Expectations
• California Consumer Privacy Act (CCPA) (opens in a new tab)California Attorney General
• UK GDPR guidance and resources (opens in a new tab)UK ICO
• Australian Privacy Principles guidelines (opens in a new tab)OAIC
• Personal Data Protection Act overview (opens in a new tab)PDPC Singapore