data model

A data model is a defined structure that describes how data is organized, named, related, and constrained within and across systems. It specifies the entities (such as orders, lots, serial numbers, equipment, and test results), the attributes of those entities, and the relationships between them.

In manufacturing and regulated operations

In industrial and regulated environments, a data model commonly refers to the way production, quality, maintenance, and business data are structured across systems such as MES, ERP, QMS, LIMS, and data historians. It typically includes:

• Core entities, for example work orders, batches/lots, serialized units, materials, equipment, and operators
• Key identifiers, such as order numbers, lot IDs, serial numbers, and equipment IDs
• Relationships and genealogy, such as which components went into which finished unit, or which tests were run on which batch
• Rules and constraints, for example one serial number per physical unit or required links between a test record and a lot
• Logical groupings used for reporting and KPIs, such as how shift, line, and product family are associated to events and measurements

A data model may be documented as diagrams, database schemas, or configuration in integration tools. It can exist at different levels of abstraction, such as:

• Conceptual data model: High-level view of the main entities and relationships, often used with business and quality stakeholders.
• Logical data model: More detailed description of attributes and relationships, independent of specific database technology.
• Physical data model: The actual implementation in databases, message schemas, or APIs used by systems.

Operational relevance

In daily operations, a clear data model supports:

• Traceability and genealogy, by defining how order, lot, and serial keys connect production events and quality records
• System integration, by aligning identifiers and relationships across MES, ERP, QMS, historians, and analytics platforms
• Reporting and KPIs, by specifying which data sources and joins underlie each calculation and how results map back to units, equipment, or time periods
• Change control, by allowing controlled updates to structures and relationships when processes or systems change

Common confusion

• Data model vs. database: A database is the implemented storage system; the data model is the design that describes what is stored and how it is related.
• Data model vs. data schema: A schema is often the concrete, technical representation (for example tables and columns). The data model includes the schema but also the conceptual definitions, rules, and intended use of the data.
• Data model vs. process model: A process model describes workflow steps and sequences of activities. A data model describes the information related to those activities and how that information is linked.

Audit and compliance context

In audits and investigations, a well-defined data model helps show how high-level metrics and reports can be traced back to underlying records. For example, it can demonstrate how KPI values are linked to specific orders, lots, serial numbers, test results, and system-of-record transactions, and how those links are preserved through integrations and data transformations.