To stop or not, that is the question: When should a buyer hit the brakes for supply base rationalization?

This paper analytically and experimentally studies the buyer’s sequential decisions in supply base rationalization during re-sourcing. Starting with a mix of qualified and potential suppliers, the buyer sequentially decides whether or not to screen another randomly selected potential supplier. Adopting the frameworks of optimal stopping problem and discrete convex analysis, we characterize structural properties and identify sufficient conditions for optimal control-limit, control-band, and one-step look-ahead policies. To investigate human decision biases, we design experiments where subjects make decisions based on the one-step look-ahead policy. Results show that compared to theoretical predictions, subjects stop with too few qualified suppliers when qualification costs are low and too many when costs are high. To explain this “too few/too many” pattern, we propose a behavioral model incorporating aversion to stopping, loss aversion, and gain-seeking biases. The structural estimation reveals that while the aversion to stopping bias exists in all treatments, the loss aversion bias only occurs in high passing probability treatments, and the gain-seeking bias in low passing probability treatments. Our findings suggest that buying companies should avoid stopping too early when the qualification passing probability is high and be cautious of betting behaviors when the probability is low.