Harmonizing product master data with retail sales for 95% accurate SKUs matching

Sigmoid helped a global consumer goods company reduce manual efforts and improve data accuracy for sales and market operations teams.

Business Scenario

A multinational consumer packaged goods firm in the household and personal care sector faced challenges with low fill rates and inconsistent POS sales data from retailers and distributors. Managing this data manually with over 50 resources was labor-intensive and error-prone. They required an automated solution to extract and standardize product attributes, eliminate duplicates, and create a single source of truth. This solution would improve accurate sales reporting, assortment optimization, and trade fund management, ensuring better investments across retailers.

Sigmoid Solution

Sigmoid built a prediction engine leveraging BART (Bidirectional and Auto-Regressive Transformer) to automatically manage and organize vast amounts of product data across different retailers. This product matching model helped in data cleaning and standardization for in-house and retailer datasets, defining and standardizing desired product attributes, developing attribute extraction models, and deduplicating the product catalog to create a single source of truth (SSOT) using the extracted attributes

Download case study

Business Impact

Sigmoid’s solution enhanced their ability to manage and organize product data on various online retail channels and clean up their MDM. The automated attribute extraction and product matching processes led to the creation of harmonized reports for sales health tracking, enabling better inventory management, trade fund management, and assortment and shelf display strategies.