Analyst Report

IDC Spotlight Report: The Power of a PX Strategy in Omnichannel Commerce

Download NowDownload NowDownload Now
Akeneo-Logo Akeneo-Logo
Prova la Free TrialProva la Free TrialProva la Free Trial DemoDemoDemo

Case Study: Rexel

As an electrical supplies distributor with more than 2.5 million products, Rexel understands the impact that high quality, reliable product information can have on the customer experience.

Unfortunately, not all of our suppliers can do ETIM. It is an international model that is not cost-free and requires significant implementation by suppliers. Following this observation, we sought to categorize our products, in the reference model [ETIM], instead of our suppliers. This action is very time consuming and requires a lot of resources and product knowledge. Based on this new observation, we looked for solution from algorithmic solutions.

Patrice Gouineau, Director of Product Data

Project Summary

About Rexel

Solutions Used

Il problema

Leader in the professional distribution of electrical supplies, Rexel has become a major player in its sector with its 2.5 million active products and its participation in the birth of FAB-DIS jointly created with the FGME (Federation of Electrical Supplies Wholesalers). Since adopting FAB-DIS, Rexel has been using all the data collected in its webshop, encouraging manufacturers to enrich the data they share. To be successful in a digital world, you have to be able to classify your product data. At Rexel, we have a historical nomenclature and an international nomenclature, called ETIM. Since the implementation of FAB-DIS, the ETIM classification is requested to our suppliers. This classification is essential for the user experience, if it is missing it is complicated to find a product on our webshop [the product will not be classified in the right category and therefore not found].

Unfortunately, not all of our suppliers can do ETIM. It is an international model that is not cost-free and requires significant implementation by suppliers. Following this observation, we sought to categorize our products, in the reference model [ETIM], instead of our suppliers. This action is very time consuming and requires a lot of resources and product knowledge. Based on this new observation, we looked for solution from algorithmic solutions.

The relationship built with Unifai is tied to our desire to have data quality. Out of over 2 million products, we have about 800,000 products that need to be completed on classification aspects.

Franck Duvigneau, Data Expertise Project Manager

La Soluzione Akeneo PIM

Four years ago, we challenged a service provider who proposed to manually enrich our products, but after a few months, we quickly understood that product knowledge was a key element to achieve this. We also challenged internal sources, with good results, and AI service providers who presented us “magic” solutions where all we had to do was press a button and everything would be done. After testing another provider, Supplier Data Manager (previously Unifai) was proposed to us by our PIM solution provider Akeneo as an AI solution knowing the ETIM model. The knowledge of the semantics and the vocabulary in which we evolve is essential for us. General algorithms without business knowledge, which were previously proposed to us, had not been able to work for this reason.”

The ambition for Rexel is for Unifai to become an everyday tool for the product data team and manual work is reserved only for the very special cases where we need an expert to validate the AI proposition.

Patrice Gouineau, Director of Product Data

I risultati

This year and especially with Unifai, we went from a 1% annual increase to 10% in categorizing our products into ETIM classes. Of our 800,000 products without an ETIM class, AI found 10% more ETIM classes than the previous year. The objective with Unifai is to have 100% of our business with ETIM classes. It is reassuring for the moment because we challenged Unifai on product data that our internal algorithm could not do and today Unifai knows how to propose classes. The goal with a tool like Unifai is that everything that is complicated is handled by the algorithm.

We tested Unifai with already corrected values to measure the gap between the in-house developed algorithm and Unifai’s. We got 98% alignment between the two algorithms and on the remaining 2%, even the human eye has trouble finding a solution.

With Akeneo PIM and AI technology powered by Unifai (now Supplier Data Manager) Rexel was able to:

  • Enrich 800,000 products to the international classification standard for technical standards, ETIM
  • Provide suggestions for content gaps or missing data with up to 98% accuracy
  • Benchmark 12,000 ETIM classifications over just 3 months

 

Sei pronto a passare alla prossima tappa?

I nostri esperti Akeneo sono qui per rispondere a tutte le domande che potresti avere sui nostri prodotti e aiutarti a progredire nel tuo percorso di PXM.