Master Data Management
Master Data Management for Dairy Business – enterprise master data platform developed for the US dairy industry to create a single, trusted view of critical business data across multiple operational systems.
Cross-functional development team consisted of
- Architect
- Project manager
- Backend developers
- Frontend developers
- Quality assurance engineer
- Data stewards
Master Data Management (MDM) is an enterprise platform for creating a single, trusted view of critical business data shared across multiple operational systems. It identifies records that refer to the same farms, haulers, and other dairy business entities, connects them into consistent golden records, and prevents duplicate or low-quality data from spreading across the organization.
CONCEPT
Master Data Management is a general enterprise approach for creating a single, trusted view of key business entities across multiple systems. In this project, the solution was applied to the US dairy industry, connecting master data across three major operational platforms: an enterprise dispatching solution for milk pickup and delivery, an enterprise payroll solution that defines how milk is paid for, and a mobile application that replaces paper manifests for truck drivers.
The three operational systems remain the systems where business data is created and stored. MDM introduces an additional master-data layer that identifies which records across those systems represent the same real-world entity and connects them through a common golden record.
Matching is defined separately for each type of entity during the discovery phase, based on the available data and its quality. Wherever possible, common unique business identifiers are used. Where such identifiers are not available or reliable, matching can be based on combinations of attributes such as name, address, ZIP code, or other business-specific data.
Once records are matched, MDM becomes the common reference point between the systems. Each application can continue using its own internal identifiers, while MDM maintains the mappings between them. Integrations that previously depended on manually maintained cross-system mappings can therefore resolve these relationships automatically from MDM data.
Changes to shared master data are also synchronized across the connected systems, removing the need to maintain the same information separately in multiple places.
A custom web application was developed for MDM oversight and administration. It is tailored to non-technical data stewards, allowing business users to review mappings, resolve questionable records, and maintain master-data quality without requiring technical expertise.
TECHSTACK
Backend
Boomi
Java
RabbitMQ
Custom Web Intefrace
Vaadin Hilla
Database
PostgreSQL, NocoDB
Build and deployment
AWS
VALUE
The solution replaces manually maintained mappings between operational systems with a centralized and automated master-data process. By creating a common reference point for business entities, it keeps entity identities and shared information consistent across dispatching, payroll, and mobile operations while reducing manual reconciliation and duplicate data maintenance.
Built-in data-quality controls identify potential duplicates and suspicious records before they can affect connected systems. Questionable data can be quarantined for review, while alerts enable data stewards to quickly resolve issues. This prevents data-quality problems from spreading across operations, integrations, and reporting, providing a more reliable foundation for business processes and decision-making.
PROCESS
The project started with a discovery phase to document the existing systems, identify the business entities and shared attributes to be governed through MDM, and define entity-specific matching rules based on available data and its quality. This was followed by solution architecture covering system components, integration flows, data ownership, security, scalability, deployment approach, and the interaction between MDM and the three operational systems.
Development was organized in two-week sprints and included integration processes, MDM-related backend capabilities in all three source systems, a custom web portal for data stewards, testing, and Infrastructure as Code. The rollout was iterative, starting with smaller and simpler entities to validate the solution in production before moving to the full migration from the legacy approach, completed within a one-hour deployment window. After go-live, the team continued with bug fixes and operational support for data stewards.
Workshop
Meetings with clients
Business analysis
Architecture
Creating diagrams
Writing documentation
Development
Integration processes
Backend
Frontend
Quality assurance
Infrastructure
Infrastructure as a Code
Support
Bug-fixing
Knowledge transition
Supporting data stewards
CONCLUSION
The MDM solution established a reliable master data layer across three critical operational systems while preserving their existing roles and processes. It replaced manual cross-system mappings with automated, governed data relationships and provided a scalable foundation for consistent data management across the dairy business.
NEXT CASE
© Deegloo. All rights reserved 2023