Dec 05, 2025 · powerbi · supply-chain · automation
LATAM Inventory Health Dashboard finally stops breaking
Built a Power BI dashboard for the LATAM supply chain team that compares safety stock vs on-hand stock and classifies each material-country pair as green / yellow / red.
Context. Daily/weekly files were being handled manually. Definitions drifted across countries. The dashboard broke the moment a column changed.
What broke. DAX logic turned into spaghetti; Incorta reports used different material-code formats; joins failed silently; automation was fragile.
Takeaway. A stable pipeline beats a fancy dashboard. Normalize, validate, and make the operational language stable first.
What was happening
The workflow failed in small ways: inconsistent columns, missing tabs, mixed currencies, and country-level “creative formatting.”
The dashboard wasn’t wrong because visuals were wrong. It was wrong because inputs were unstable.
The real problem
The dashboard was acting like a data pipeline, but it had no pipeline discipline.
No contracts. No validation. No stable definition of what a row means across countries.
What I changed
Standardized intake and semantics so green/yellow/red means the same thing every time.
Made the model more robust to upstream drift instead of letting Power BI guess.
What I’d do next
Add basic data quality flags and ingestion history.
Add weekly roll-ups: what flipped status and what stays chronically red.