Loading...

Media is loading
 

Description

In obtaining low-cost goods, the indirect expenses associated with sourcing suppliers can be substantial compared to the potential advantages of lower direct purchase costs. We addressed this problem as an "exploration" vs. "exploitation" trade-off. The proposed methodology uses a Bayesian technique to learn a stochastically optimal sourcing strategy directly from quotation data. We illustrate our approach using real quotation data for the procurement of electronic resistors (n=201,187). Rather than making optimal predictions, we concentrate on making optimal decisions. In doing so, we offered a significant improvement in purchase and procurement process costs. Our model is also more robust to prediction errors.

Share

COinS
 
Jan 17th, 12:00 AM

Prescriptive Analytics in Procurement: Reducing Process Costs

In obtaining low-cost goods, the indirect expenses associated with sourcing suppliers can be substantial compared to the potential advantages of lower direct purchase costs. We addressed this problem as an "exploration" vs. "exploitation" trade-off. The proposed methodology uses a Bayesian technique to learn a stochastically optimal sourcing strategy directly from quotation data. We illustrate our approach using real quotation data for the procurement of electronic resistors (n=201,187). Rather than making optimal predictions, we concentrate on making optimal decisions. In doing so, we offered a significant improvement in purchase and procurement process costs. Our model is also more robust to prediction errors.