Case Study

Rapid solution for optimized demand forecasting at the most granular level

Published on 05th January 2022 | By Mate Labs

For

A consumer goods giant

Tags

Demand Forecasting
Rapid Forecasting
Cost Reduction

Challenges

Before Mate Labs’s solution was implemented, their corporate client used an established legacy tool for demand forecasting that lacked at-scale, advanced analysis capabilities, and failed to measure the impact of promotion and other dynamic features. This also meant a huge wastage of resource time and effort in finalizing demand planning.

Solutions

Mate Labs’s AutoML platform, Mateverse successfully captured the hidden patterns in forecasting demand via automated feature engineering. In the concerned use case, Mateverse’s feature engineering unpacked and analyzed the impacts of the following features to boost forecast accuracy:

  • Moving Average of Actuals
  • Actual Fill-rate
  • Events data
  • Promotion data
  • Features at different granularity i.e. ASM, Depot, Chain Month, Year, ASM, Depot, Portfolio as Features
  • Seasonal SKUs

  • Key Metrics

    15%

    15% jump in forecast accuracy (a total of ~75%) over their traditional method of forecasting which only gave them ~60% accuracy

    $530k

    $530k cost-savings from the forecast accuracy achieved

    Similar Case Studies

    For
    Procedure digitization
    Machine downtime
    Manufacturing
    Connected machines
    Automotive
    For
    Industrial automation
    Real time monitoring
    Manufacturing
    Machine utilization
    Automotive
    For
    Improved productivity
    Quality check
    Manufacturing
    Defect Detection
    Visual Quality Inspection