Our customer was having trouble with procurement due to rapid changes in market preferences and fluctuations. Although they could quickly source components, they, too often, were buying the wrong types, leading to shortages, wasted inventory, and unnecessary surplus. The result was spiraling procurement costs and delays in the company’s production and supply chains. The company’s procurement models are built on historic data, as opposed to more dynamic data streams.
By tapping into thousands of external data sources that offered greater context, our customer was able to build a more reliable forecasting model that optimized predictions. The new model identified key features to determine demand, including:
Switching to an ML-based and dynamic demand forecasting system allowed the manufacturer to boost its accuracy by 9% in the short run, which led to a rapid reduction in inventory costs of nearly 17%. More importantly, it enabled the company to better adapt to rapidly changing manufacturing conditions. The company overhauled their procurement process to just-in-time shipping, empowered by accurate forecasting that reduces costs without sacrificing quality.