
ARTICLE SUMMARY
- Collecting machine data is not the same as understanding production. A data foundation connects readings with the equipment, products, and work they describe.
- Consistent definitions, accurate timing, and clear ownership help manufacturers build production information their teams can trust.
- Clear data ownership helps manufacturers resolve conflicting records, maintain consistent definitions, and keep production information trustworthy as operations change.
- Rain Engineering helps manufacturers connect equipment and systems while aligning manufacturing information with their operational needs.
Having production data is not the same as having a clear picture of production.
Before your team can trust a report, it needs to know where the numbers came from, what they represent, and whether they describe the same work. A manufacturing execution system, or MES, depends on those answers.
Its data foundation is the combination of connections, definitions, and responsibilities that makes manufacturing information reliable and useful.
Industrial data modeling and governance provide the structure behind that understanding.
Today we take a closer look at what an MES data foundation includes, how it supports better production decisions, and where manufacturers should begin.
Why Collecting Data Is Not Enough

An MES helps manufacturers manage and track shop-floor production. It uses information about equipment, materials, and work activities to support production execution and visibility.
Connecting equipment is necessary, but connectivity answers only one question: can the system receive information? It does not establish whether the information is complete, correctly interpreted, or associated with the right production activity.
A machine reading needs an explanation. The receiving system must know which equipment produced it, what was measured, and how that measurement relates to the work underway. Industrial data modeling establishes those relationships so applications can use raw signals as meaningful production information.
The distinction is simple. Collecting data tells you what information is available. Building a foundation tells you what that information means and whether you can rely on it.
For the customer, the objective is not a larger collection of numbers. It is a clearer understanding of production.
What Makes Production Information Trustworthy?
Trustworthy information answers a few basic questions: where did it come from, what does it mean, when did it happen, and which work does it describe?
The first two questions require consistent identities and definitions. Equipment and measurements need recognizable names, but they also need documented meanings. Teams should agree on units, collection methods, and the difference between measured, calculated, and manually entered information. These standards help applications interpret data consistently.
A tag identifies a data point in an industrial system. However, the tag name is only a label. Its definition explains what someone should understand when they read the value.
Timing matters too. Information needs a trustworthy record of when the source reported it. It also needs an indication of whether the value is usable. OPC UA, an industrial communication standard, supports values accompanied by timestamps and status information.
These details help the team distinguish a valid reading from information affected by a communication or source problem. A number on a screen should not automatically be treated as a trustworthy account of current production.
Connect the Numbers to the Work

Knowing what a machine measured is only part of understanding production. The team also needs to know which product was running, which order was active, and which equipment performed the work.
Those relationships are called production context. They connect machine information with orders, products, batches, materials, and process steps. Context turns an equipment observation into information about a particular production activity.
This is also why shared records matter. Products, materials, and equipment need consistent identities across connected systems. These shared records are commonly called master data.
When applications disagree about those identities, an interface can move information successfully while still producing inconsistent records. Master-data management addresses that problem by establishing authoritative records and controlling how they are maintained.
The practical benefit is shared interpretation. People reviewing the information should not have to guess whether two records refer to the same material, machine, or production activity.
Start With the Decision, Not the Technology

The best starting question is not, “How many tags can we collect?” It is, “What does our team need to understand or manage?”
Manufacturing data guidance recommends beginning with business needs, examining existing systems, and establishing information standards before expanding the architecture.
Define the production decisions the MES must support. Then identify the information required, where it currently exists, and what prevents the team from using it confidently.
That assessment helps distinguish different problems. A measurement may genuinely be missing. Alternatively, the measurement may already exist but lack a clear definition or connection to production.
More sensors address missing measurements. Better organization addresses missing meaning. The assessment should determine which problem the plant actually has before selecting the solution.
Begin with a manageable production scope. Establish what reliable information should look like, verify it with the people who understand the process, and use that work to guide expansion.
Give Someone Responsibility for the Information

A foundation needs maintenance. Equipment changes, product records evolve, and production procedures develop.
Someone must decide who maintains definitions, who approves changes, and who resolves conflicting records. This responsibility is called data governance. It covers the accuracy, consistency, security, and usability of information throughout its lifecycle.
Governance does not mean one person owns every detail. It means responsibilities are clear enough that problems have an owner.
Rain Engineering’s MES implementation services include assessing existing systems, mapping integration requirements, aligning business objectives, and connecting Proficy Smart Factory MES with ERP, SCADA, and shop-floor systems. That work connects the technical project with the information the operation needs.
The central message is trust. A useful MES foundation helps your people understand where information comes from, interpret it consistently, and know who maintains it.
Those principles are straightforward. The technical work is making them consistent across equipment, interfaces, and applications.
(*NOTE: For your convenience, this article has been divided into two portions: a simplified introduction explaining what a data foundation is and why it matters – above -, followed by a technical deep dive for readers who want to understand how it is built – below.)
The Technical Deep Dive
Now that we have covered the essentials, we will take a deeper look at the technical aspects of building an MES data foundation, including equipment models, tag definitions, timestamps, data quality, and system integration.
Build the Equipment Model Before Mapping Individual Signals

An asset model establishes how production equipment is organized and related. It gives connected applications a shared structure rather than allowing each system to invent its own equipment identities.
ISA-95-based modeling provides a framework for relating enterprise, site, and production assets. The implementation should reflect actual operating relationships while maintaining consistent identities.
Keep the equipment identity separate from its technical connection wherever the architecture supports that distinction. An equipment record describes the production asset. A controller address or tag path describes how to access information from it.
The associated tag dictionary should document source, asset, data type, engineering units, meaning, and collection behavior. Include scaling and transformations so the reported value can be traced back to its source.
The acceptance question is practical: can the team identify what each value represents and verify how it was produced?
Preserve Time and Quality Through the Interface
The OPC UA DataValue specification distinguishes the value itself from the information needed to interpret it. Its structure includes a status code, source timestamp, and server timestamp.
The source timestamp represents the time assigned by the data source. The server timestamp represents when the server received the value or knew it to be accurate. These are different observations and should not be treated as interchangeable.
OPC UA specifies that source timestamps use UTC and should be generated as close to the source as possible. A server receiving a value from another OPC UA server must pass that source timestamp through unchanged.
Status codes communicate good, uncertain, or bad conditions. The integration should preserve those conditions and establish how consuming applications respond.
Define acceptance criteria for both timing and quality. Verify that the receiving application retains the intended timestamps and does not silently convert unavailable or invalid information into an apparently valid measurement.
Define Events and Their Production Relationships

Measurements describe conditions. Production events identify meaningful occurrences or transitions.
The implementation must specify which events the MES requires, how each event is detected, and which production identifiers accompany it. Event modeling connects machine behavior with the operational context needed by downstream applications.
For each event, document its trigger, equipment identity, timestamp, and required production relationships. Establish how the system determines the active order, product, batch, or process step.
Review these definitions with operations before implementing the interface. A technically convenient trigger is not automatically the correct representation of how the plant records production.
Validation should establish whether events appear at the correct boundary and belong to the correct work. Confirm the resulting records against the approved operating definitions, not merely against whether a message arrived.
Assign Master-Data Ownership and Change Rules

Master-data ownership determines which system maintains each shared record and which systems consume it.
Document authoritative sources for product, material, equipment, and production-definition records. Establish how changes are approved, distributed, and made effective. Governance and master-data management address both shared definitions and responsibility for maintaining them.
Also define how interfaces handle records they cannot recognize. Missing identifiers or conflicting versions should produce an actionable exception rather than an unexplained result.
Avoid allowing integration code to become the undocumented owner of production rules. Where a transformation is necessary, record its purpose and identify who approves changes.
The acceptance question is whether connected applications use the intended identity and definition at the intended time. Successful record transfer is only one part of that test.
Create a Data Contract, Then Choose the Architecture

A data contract describes what an interface exchanges and how each field should be interpreted.
It should specify required identifiers, units, timestamps, quality conditions, event definitions, and validation responsibilities. Shared modeling and governance provide the basis for those agreements.
The contract should also address failure behavior. Define what happens when information is incomplete, late, repeated, or rejected. Verify buffering and replay capabilities where available rather than assuming that a restored connection creates a complete history.
A unified namespace, or UNS, can organize shared manufacturing information for multiple applications. It is one architectural approach to connecting operational sources and context, not a replacement for defining them.
Choose communication and storage technologies around the information requirements. A protocol can deliver a field correctly while two applications still interpret it differently. Transport and meaning therefore require separate, coordinated validation.
Use the Historian for Its Intended Role
An industrial historian collects and stores time-series information and events. An MES supports production execution and related operational records. These capabilities complement one another, but neither automatically supplies every element of the data foundation.
GE Aerospace provides a documented example of shared industrial data infrastructure. Velotic’s Proficy Historian product page reports that GE Aerospace manages operational technology data across 32 manufacturing plants and moved from 32 separate deployments to a single deployment on AWS.
The relevant point is not the plant count or hosting platform. It is the value of managing industrial information consistently. Velotic also identifies standardized data formats as a way to simplify comparison across sites.
The example illustrates shared infrastructure and standardization. It does not establish that every manufacturer needs the same architecture or that historian deployment alone creates trustworthy MES records.
Validate the Information and Keep It Maintained

Validation should prove operational meaning as well as connectivity.
Compare received values with their sources. Verify transformations, timestamps, quality conditions, and production associations. Reconcile records against approved reference information and investigate differences.
Test interruption and recovery behavior through an approved plan. Document gaps so users understand the limits of the information they are reviewing.
After commissioning, maintain definitions, ownership assignments, and validation rules. Governance remains an ongoing responsibility.
A strong MES data foundation is not simply a technical layer beneath the software… It is the shared understanding that connects equipment activity with production decisions.
Reliable measurements, consistent identities, meaningful context, and clear ownership give manufacturers a better basis for using the information they already collect. The goal is not more data for its own sake. It is information the team can understand, maintain, and trust.
FAQ
Q: What Is a Data Foundation for MES?
A: It combines connections, definitions, production context, quality controls, and ownership to make manufacturing information reliable and usable.
Q: Do We Need More Sensors?
A: Not necessarily. First determine what the MES must support and whether existing information meets that need when properly organized and contextualized.
Q: Is a Historian Enough?
A: No. A historian provides industrial data collection and storage. Shared definitions, production relationships, and ownership are also required.
Q: Who Maintains the Foundation?
A: Responsibilities should be assigned to the functions that maintain equipment, production definitions, quality requirements, and business records. Governance establishes those decision rights.
P.S. Your MES should receive information that reflects how your plant actually operates. Rain Engineering provides Proficy MES integration services that connect equipment and existing systems with manufacturing and enterprise applications. Its assessment and integration work helps align those connections with operational requirements.
Contact Rain Engineering to discuss the information your MES needs and how to build a foundation your team can trust.

