Coordinating supply chains

I just attended a talk given by Tava Olsen from Washington University in St. Lous. The talk was held at the business school at the University of Auckland (who are building a spectacular new building nearby). Her talk was an overview talk on supply chain coordination. For a non-specialist like myself (albeit one who often attends supply-chain-oriented conferences), it was really useful to get an overview of how research in the area of supply chain coordination works.

Here is Tava’s “Make your own Coordination Paper” recipe for research in this area (as I interpret it):

1. Find a supply chain and identify the actors
2. Determine the optimal decentralized actions for the actors
3. Determine the optimal centralized result (perhaps by pretending all the actors were one)
4. Find an implementable contract so that the decentralized actions give the centralized result.

The example she gave was a simple two stage supply chain with a retailer ordering from a supplier in a single period, one order environment. The retailer’s problem is a simple newsvendor problem, with the amount ordered limited by the losses due to possible overordering. But if the supply chain was centrailized (think of the retailer and the supplier owned by the same company), then the loss of overordering is less (it is the production cost rather than the higher wholesale cost), so more is ordered. The result is more expected profit in the supply chain. But, crucially, that higher expected profit occurs at the supplier, not the retailer who does the ordering. How can the supply chain be coordinated so that the pair find the higher profits?

There are lots of possible solutions: a fixed payment from the supplier to the retailer for ordering the higher quantity would do, but that is rather inelegant (there are few contracts written of the form: “I’ll pay you $1000 for ordering exactly 621 items”) and requires a huge amount of information on the part of the supplier. “Better” solutions (more implementable, while still coordinating) might involve “buyback” contracts where the supplier agrees to take back unsold inventory at a particular value or quantity flexibility contracts, where the supplier agrees to provide any amount within a range of values.

Tava continued with a very nice example of supply chains like those between automobile manufacturers and suppliers whereby parts absolutely must be provided on time (penalties of thousands of dollars per minute late are common). In such cases, suppliers must expedite parts in the case of possible shortages, which can be a very expensive process. Coordinating across such a supply chain means having the manufacturer make decisions that keep in mind the expediting costs.

I was very happy to get such an overview talk: it put a lot of supply chain research into perspective.