How to Lessen Retest Downtime in Test Process?

Ever heard the one about the Japanese company that received an order for 1000 circuit boards with less than 0.3% duds? The infuriated Sales Director shipped 997 in one large package and 3 in a smaller one. When the Goods Inwards Clerk made enquiry about the way that the order had been shipped, the response was "The smaller box contains the three dud ones you ordered".

The nature of the project must be, therefore, why is this particular line making duds that need rework, retest or whatever, and what can be done to eliminate the making of those duds in the first place <rhetorical question>?

Consult any textbook on Japanese manufacturing techniques. The operation described in the original posting has much to learn from it - preferably before it goes out of business.

Nothing to add, except GA.