SAP Master Data Quality Programme
The unglamorous project that made everything else possible: product and business partner master data taken from 70% of records carrying issues to 2%, in one year.
Role
Data Scientist
Timeline
2024 to 2025
The challenge
70% of SAP master data records carried issues, and every one of them compounded downstream: wrong reports, broken processes, rework in every department that touched the data. No analytics or AI initiative could be trusted on top of it.
The decision
Chose Fix at the point of entry over Clean downstream
Downstream cleaning is a treadmill: the same errors return every month. Governance rules where data is born stop the problem from existing.
How I approached it
- 01Audited the master data systematically to find not just the errors, but the entry points and process gaps that produced them.
- 02Traced root causes across departments: most bad data wasn't carelessness, it was missing rules about who enters what, when, and in which format.
- 03Built governance rules and quality checks at the point of entry, with clear ownership per data domain.
- 04Tracked the error rate monthly and made it visible, so quality became a number departments watched.
Impact
- ✦Master data issues from 70% of records to 2% in one year.
- ✦Downstream errors and rework cut across every consuming system.
- ✦Reports and processes inherit clean data instead of compounding bad data.
- ✦The trust foundation that made BI and the AI portfolio credible.
Stack
- SQL Server MDS
- SAP
- SQL
- Power BI