An article in the Wall Street Journal (May 21, 2012) describe an approach called twinning, pairing a weaker and stronger plant to share productivity improvement methods, and thus increase productivity. ArcelorMittal paired their Burns Harbor plant in Indiana with a highly automated facility in Gent, Belgium. The goal was for Burns Harbor to implement techniques from Gent to match its productivity. This required employee exchange, shifting from pencil to computer, computerized poruing of liquid iron etc. The result – Burns Harbor is at 1.32 man hours per ton, second only to Gent at 1.25 manhours per ton and more efficient than the US average of 2 manhours per ton. How does twinning speed up productivity improvements across plants, and should it be accompanied by appropriate incentives to be successful ? Given that the better plant will continue to improve, does twinning create a continual race to be better and thus help both plants ? Given that the productivity improvement at Burns Harbor had to be accompanied by headcount reduction to increase profitability, how should the improvement process be sustained ?