Heineken-Alken Maes - Turning fuzzy data into one clean view
Every wholesaler sent Alken Maes their sales data in a different format — impossible to compare, let alone analyze. We built the fuzzy matching that turned it into one standardized, analyzable dataset.
The challenge
Alken Maes sells beer to wholesalers, but once that stock left the warehouse, they lost visibility into who bought it next: the bars, restaurants and hotels where it's actually consumed and sold. To close that gap, they set up an incentive program — wholesalers who shared their sales data would receive a discount. But every wholesaler exported that data in its own format: different columns, different units, different ways of describing the same products. None of it could reliably be compared or analyzed as it was.
The solution
We built a system in Python that standardizes every incoming file into one common structure. Fuzzy matching logic reconciles product descriptions that differ from wholesaler to wholesaler — the same case of beer described a dozen different ways — into a single, unified product and client view.
Alken Maes could see exactly what they sold to wholesalers, but nothing about where it went from there. Every wholesaler sent us their sales data in a different CSV format — different columns, different product names for the exact same case of beer. We built the fuzzy matching that pulled it all into one standardized view, so for the first time they could see the full chain, from warehouse to bar.
Results
- Standardized 6 different wholesaler ERP/export formats into one structure
- Replaced a manual, once-a-month matching process with no certainty scoring with a repeatable, automated one
Services delivered
- Data Standardization
- Fuzzy Matching
- Python
- Product Data Unification
