Dual Sourcing: What a Second Supplier Protects

How to read this sheet

What a second source protects is not that it exists but what it could deliver in the weeks after the first one stops, which is its normal volume multiplied by whatever it can lift to. A supplier on a tenth of the volume with a surge of two covers a fifth of the requirement, and the other four fifths are exposed exactly as they were before it was qualified. Full cover therefore requires the other suppliers to be holding one over the surge multiple of the volume already - more than half of it at any realistic surge - which almost no arrangement called dual sourced does. Read the expected-value verdict carefully. On most books the premium exceeds the expected loss it removes, and that is not an argument for single sourcing: firms split supply to cut the tail rather than the mean, and an expectation cannot see a tail. The failure probability reported is the one at which the mean alone would justify the spend, and the honest use of it is to ask whether you believe the real rate is above or below it - not to treat the arithmetic as the decision. Several things are outside this. Correlation is the largest: a second source in the same region, on the same raw material or sharing a shipping lane is not independent of the first, and this sheet treats every disruption as unrelated to every other, which overstates the protection. Nor does it model partial disruption, the loss of price leverage from concentrating volume, the quality and shade differences between two mills running the same specification, or the fact that a supplier held at a token share may not answer the phone in a crisis - which is the practical argument for a larger share and one this arithmetic reaches by a different route.