Clean the master data that execution depends on

Prepare MES master data by assigning ownership, defining meanings and testing how changes reach execution. Loading a spreadsheet successfully does not prove that item codes, units, routings and resource definitions are suitable for controlled production work.

Inventory the objects execution needs

Start with the selected workflow. Identify items, material lots, units of measure, operations, routings, resources, instruction references and status codes. Do not import every available ERP field simply because it exists. Each object should have a purpose, an owner and a known source.

Distinguish master data from transactional state. An item definition differs from the current status of a particular material lot. A routing template differs from the approved revision used for one order. Confusing those concepts can produce records that look plausible but cannot explain what instructions or constraints applied at the time work occurred. Document the distinction before building mappings.

Define authoritative ownership and change paths

Object	Question to settle
Item	Who creates, revises and retires it?
Unit	Where are conversions defined and approved?
Routing	How does a revision become effective for an order?
Resource	How is a machine or workstation identified consistently?
Status	Which transitions are permitted and who authorizes them?
Instruction	How is the approved version linked to execution?

The table is a starting point, not a prescribed enterprise model. If MES adds local attributes, specify which ones it owns and how they relate to upstream data. Avoid uncontrolled copies that drift independently. A named owner should be able to explain how a correction is approved and propagated.

Test units and quantities with deliberate variation

Use examples that expose unit differences: pieces versus cases, pounds versus kilograms, or a batch quantity versus an individual unit. Define conversion precision and rounding where relevant. Do not assume a field named quantity has the same meaning in every system. Include scrap, partial completion and rework if the selected workflow uses them.

An illustrative test might issue two containers recorded as 25 pounds each and verify that the receiving system interprets the total as 50 pounds, not two pounds or 50 kilograms. The example is hypothetical. The plant must establish its actual units, tolerances and controls, particularly where quantity errors can affect quality or safety.

Control effective versions

Ask what happens when a routing or instruction changes after an order is released. Does the order keep its approved version, require an authorized update or stop for review? The correct rule depends on the process and applicable requirements. Make it explicit and test it in the intended configuration.

Preserve enough version context to interpret historical records. A report that always joins history to the latest master may misrepresent what was used at execution time. Test retrieval of an old order after changing the master. Confirm that the record can still explain the relevant instruction, operation and material meaning without relying on someone’s memory of the prior configuration.

Clean data through rules and ownership

Profile the pilot dataset for missing identifiers, duplicate codes, inactive objects, inconsistent units and unsupported statuses. Assign exceptions to the business owner. Do not silently repair values in a migration script without a documented rule and approval. Otherwise the project can create a second, unofficial source of truth.

Keep the original input, transformation rules and reconciliation results. Use a controlled repeatable load so the team can distinguish a data defect from an import defect. If a rule changes, rerun the affected checks and record the impact. A successful row count is useful but insufficient: verify relationships and meanings, not just how many records were imported.

Rehearse the cutover and ongoing maintenance

Decide when source changes stop, how late changes are captured and who authorizes the final load. Reconcile selected objects and orders after the load. Include a rollback or correction procedure appropriate to the project. Do not assume data preparation ends at go-live; new products and resources will continue to appear.

Create a maintenance procedure with owners, approval, effective dates and monitoring for failed propagation. Include the master-data process in support handover and regression tests. The pilot should demonstrate at least one controlled change from source approval through workstation use and historical retrieval. That evidence is more informative than a one-time spreadsheet import completed by the project team.

Take this to the working meeting

Bring one item, its routing, its units and a proposed revision to a working session. Have each owner explain where the object changes and when the change becomes effective. Follow the approved revision into a test order and then retrieve an older order. If the two histories cannot be distinguished, resolve that design question before importing more data or expanding the pilot dataset.

FAQ

Who should clean master data?

Business owners should approve meanings and corrections. Technical staff can profile and transform data, but should not invent business rules.

Is a successful import enough?

No. Test relationships, units, versions and the behavior of representative orders.

Should history use the newest master?

Not automatically. Preserve the context needed to understand what applied when work occurred, under the organization’s record requirements.

Continue your decision

Use the pilot readiness to document your own inputs. Compare operating models on the comparison page.

Sources and scope

NIST Smart Manufacturing — Official research context for manufacturing systems. The decision worksheets in this guide are editorial recommendations, not NIST requirements.

Examples are hypothetical and are not measured customer results, supplier quotations, or evidence of professional review. Prepared on October 30, 2026; confirm current requirements with the relevant official body and your responsible specialists.