Img-Reboot

Used Sancus to help a leading industrial company get better visibility into revenue and sales opportunity by product category

Summary

Client is the largest U.S distributor of plumbing supplies, PVF, waterworks, fire and fabrication products.

The firm had an inventory of ~3MM SKUs which had to be classified into a product family. Only ~330K out of 3MM SKUs had been classified manually by a third-party contractor at a throughput of 7K SKUs/month.

This approach was time-consuming and inefficient and warranted the implementation of an automated system that would increase efficiency.

Approach

To achieve this, we leveraged Sancus in the following aspects:

  • Data Engineering: Processed, cleansed and passed a high volume of data for approximately 3 million SKUs through the text mining pipeline
  • Neural Networks: Developed an ML algorithm using elements of supervised and unsupervised learning to classify the remaining SKUs based on existing classifications
  • Deployment: This ML based product classification solution was implemented on the cloud using Microsoft Azure

Key Benefits

The solution allowed the client to achieve product to category classification at scale with higher accuracies, providing better insights into revenue and sales opportunity

Results

Icon Boost

The monthly classification throughput increased by 28x and the total accuracy of product classification shot up to 95%.

Talk To Us