Supply Chain · Digital

Master Data Quality

Clean the basics first: item masters, lead times, bills of material, supplier records — automation amplifies whatever data you feed it.

Time45 min
FormatTeam
StageDigital

Master Data Quality: what it is and why it works

Master data quality work cleans and then protects the reference data that planning and execution systems depend on: item masters, units of measure, lead times, lot sizes, bills of material, routings, supplier records and customer ship-to data. The method profiles each master to count duplicates, missing fields, invalid values and stale records, fixes the three most damaging defect types first, assigns clear ownership for creating, validating and correcting records, and adds automated checks at entry so new errors are blocked rather than cleaned up later.

It works because planning systems, integrations and analytics amplify whatever data they receive. A wrong lead time silently shifts every purchase order for that item; a duplicate item splits stock and demand in two. Systematic profiling reveals patterns that ad hoc fixes miss, and ownership prevents the same defects from returning. It beats one-off cleanup projects because entry controls stop the decay. Master data quality underpins EDI/API Integration, since messages fail on mismatched codes, is a prerequisite for any ML Forecasting Pilot, and makes segmentations such as ABC-XYZ analysis trustworthy.

What you need

Extracts of item, supplier, bill of material, routing and customer masters
Transaction data to compare with master values, such as actual versus planned lead times
A list of known data-related incidents: wrong orders, stockouts, invoice disputes
Current data entry process and roles
Business rules defining valid values for key fields

What you get

A data quality profile with defect counts by master and field
Corrected records for the top three defect types
A data ownership matrix: who creates, validates and corrects each master
Automated validation rules at data entry
A recurring data quality metric, such as defect rate per 1,000 records

When to use it

When the system’s answers are distrusted because the inputs are quietly wrong.

How to do it, step by step

Profile the masters: items, suppliers, BOMs, lead times.
Count duplicates, missing fields, stale records.
Fix the top three defect types first.
Assign ownership: who enters, who validates, who corrects.
Automate checks at entry so quality is enforced, not cleaned up later.

Worked example: Fixing planning lead times at a valve manufacturer

Illustrative scenario — figures are realistic but not from a real company.

A manufacturer of industrial valves has about 22,000 active purchased items. Planners routinely overrode system recommendations because they did not trust them. Buyers complained of constant expediting while inventory sat at record levels.

The planning manager profiled the item master and found 3,100 items with no planned lead time or a default of 1 day, about 400 duplicate items with slightly different descriptions, and 1,800 items whose planned lead time differed from the last 12 months of actual receipts by more than 50%.
The team fixed these three defect types first: lead times were reset from actual receipt history reviewed by buyers, duplicates were merged with stock and open orders consolidated, and missing fields were filled.
Ownership was defined: engineering creates items, purchasing owns lead times and suppliers, planning validates order parameters.
Entry rules were added in the ERP: no item could be released without a lead time, and a lead time change above 30% required a reason.

Result. Six months later, planner overrides on purchase recommendations fell by about half, expedite requests dropped by roughly a third, and inventory began declining without service loss. The lesson: the system had been right all along about its data; the data was wrong.

Open it in the FieldKit Supply Chain app → Get the Supply Chain template pack — join the waitlist

Common pitfalls and how to avoid them

Launching a massive cleanup of every field.Profile first, then fix the few defect types with the biggest operational impact.
Cleaning data without fixing how it is created.Add entry validation and ownership so defects are prevented, not recleaned every year.
Leaving ownership to IT.Assign each master and key field to a business owner who understands its meaning.
Measuring completeness only.Check accuracy against reality as well, for example planned versus actual lead times.

Frequently asked questions

What is master data in supply chain?

Master data is the relatively stable reference information that transactions use: items with their units, lead times and planning parameters, bills of material, routings, suppliers, customers and locations. Transactions such as orders and receipts change daily; master data defines how they are interpreted and planned.

What are the dimensions of data quality?

Commonly used dimensions are completeness, accuracy, consistency across systems, validity against rules and formats, uniqueness without duplicates, and timeliness. For supply chain data, accuracy against physical reality, such as actual lead times and dimensions, is often the most important and the most neglected.

Who should own master data?

Business functions that understand the meaning of the data should own it: engineering for bills of material, purchasing for supplier and lead time data, planning for order parameters, sales operations for customer data. IT or a data team provides tools, rules and governance support, but ownership stays with the business.

Origin

Master data management practice — data governance lineage, 2000s.

Used in these playbooks

Digital quick wins month 1 month

A month of pragmatic digitalization: clean masters, automate one document flow, stand up a lightweight tower, then pilot ML.

Related methods

More in “Digital”

Instrument the chain: data, integration and decision support.