Prediction Modeling of Demand Arrival for Spare Parts

Business Problem Statement

An American Supplier of Scientific Instrumentation, wanted to predict the scientific equipment’s Mean Time Between Failures [MTBF] to predict the demand arrival of spare parts.

Objectives

To Reduce the E&O Inventory by predicting the Demand Arrival of Spare Parts and to Reduce Customer’s Wait Time for MTO Spare Parts during the equipment failure.

Business Challenges

High E&O

Inventory Reduction

Order Service Time

Solution Overview

We replaced the legacy forecasting system with a low cost ML Model to predict the demand arrival of the spare parts. The deviation between actual & predicted in the former model was +145/-145 days with an overall order coverage of 22% in a Quarter percentage. This was replaced by novel model giving a higher accuracy and converge of 54% in a Quarter

Solution Flow

Business Outcome

Demand Arrival Prediction Helped To.

Determine the optimal inventory level to satisfy demand while minimizing stock
Order fulfillment based upon quarterly projection to achieve minimal E&O

65% of

Prediction Accuracy

for MTO SKUAchieved

Achieved 83% of

Prediction Accuracy

for the Entire SKU list

Inventory Prediction Helped To.

Maintain the right level of supplies for the rolling 6 months which resulted in lower investment costs and less waste due to overproduction or understocking
Reduce the Excess and Obsolete inventory by forecasting the demand for the rolling 6 months
Reduce the customer’s wait time by identifying the potential problems before they occur which resulted in Increased Customer Satisfaction