Case: Sonepar

Sonepar is an independent family business of 145 operating companies in 48 countries with worldwide market leadership in B2B distribution of electrical products, solutions and related services. Worldwide, Sonepar has one million customers, 48,000 employees and 3,000 branches.

Challenge

Squadra Machine Learning Company had previously developed smart solutions for Sonepar, namely:

• Automatic classification of products to the right ETIM classes, based on manufacturer and product descriptions.
• Matching of supplier feature names and supplier feature values to the ETIM features and values.
• Extracting feature values from a product description.

Now, Sonepar Asia Pacific (APAC) had identified a number of other challenges that they face with regard to the processing of data they receive from suppliers: the received data is unstructured, needs a lot of manual processing for which no resources are available, and each process is repeated in every operating company instead of being handled centrally. As a result, wrong data is sold to suppliers locally without a data governance framework in place. In order to be able to monetize vendor product data, there is a strong need for a smart online solution on the short term. 

The existing software solutions that Squadra Machine Learning Company already implemented at Sonepar were used as a basis to create a solution as a worldwide data service from the standard procurement system (SPS) to regions like Sonepar APAC. The automatic classification with Powerconvert.ai was already available for the SPS but was merely trained with a European dataset. Sonepar has therefore requested to perform a Proof of Concept (POC) project which demonstrated the applicability of the Powerconvert.ai solutions when trained with a Sonepar APAC dataset.

Solution

The existing solutions were configured and trained with Sonepar APAC data and a sample supplier product data file was processed in the English language in order to create an output file that was converted to the Sonepar ETIM structure and format. This proof-of-concept demonstrated that feature and value matching with Powerconvert.ai proposed a mapping between the suppliers features and values to the Sonepar ETIM model. The proof-of-concept included a user interface to manually change or approve the proposed mappings. No additional data was required from Sonepar for this.

The output was an excel file which contained the products from the suppliers input file, but then in the Sonepar ETIM structures and format.

Result

Now, with the help of Powerconvert.ai, Sonepar is able to convert its suppliers features and values to the brands’ ETIM data model. The proof-of-concept appeared to have a user-friendly interface which allowed Sonepar’s data specialists to manually change or approve the proposed mappings. This facilitated convenience for the brand and saved them a lot of valuable time and money.