Identified Manufacturing Bottleneck to Prevent Downtime

Issue:

A large home goods supplier noticed one of their production plants had an issue with downtime occurring more often than expected. The plant operators had hypotheses about which were the problematic products and materials. However, no proof was available to validate this as downtime fluctuated daily, weekly, monthly and on different production lines.

The production plant needed help to test the hypothesis and validate if there was a statistical difference in downtime caused by specific materials and/or products before taking corrective action.


Solution:

Collected, cleaned and connected several years of historical data from multiple databases for analysis. Analyzed downtime and conducted hypotheses testing against different product and production attributes such as raw material blends, quality checks, facility temperature and humidity to uncover and address the root causes.

Further validated trend differences by running real-time experiments on manufacturing lines and analyzing this data to isolate the conditions and raw materials most likely to prevent downtime.


Impact:

Freed up over $1M worth of manufacturing time by reducing several hours of downtime per month.