Normalizing Coupa Exports When Supplier Codes Drift Across Business Units

Drifting supplier codes across regional business units break category dashboards until you apply a governed mapping layer—not one-off Excel fixes.

Procurement analyst comparing supplier records across multiple spreadsheets

Multi-unit Korean enterprises often deploy Coupa regionally while supplier master data remains decentralized. The result: three codes for the same logistics provider, duplicate entries after M&A, and PO lines that resist category roll-up.

Step 1: Extract the full supplier universe

Pull 24 months of invoice lines with supplier ID, legal name, tax ID where available, and business unit. Do not filter inactive suppliers yet—you need the ghosts to build the map.

Step 2: Cluster by tax ID first, name second

Exact tax ID matches are your anchor. For remaining orphans, use normalized legal name matching with manual review for homonyms. We maintain a review queue spreadsheet where category managers approve clusters above KRW 50M annual spend.

Step 3: Publish a governed override table

Store mappings in a versioned table with effective dates—not hidden formulas. When a regional unit onboards a new supplier code, they submit a mapping request rather than patching the dashboard.

What this prevents

Without this layer, supplier concentration charts understate dependency and savings attribution double-counts when the same vendor appears under two codes. Fixing this after dashboard delivery costs more than building the map upfront.