What a Machine That Can Do Two Things Is Worth

How to read this sheet

The whole tool is one comparison: the best average against the average of the best. Committing to an article means taking its expected margin whatever happens; keeping the option means taking the best article in each market as it arrives. The second is always at least as large as the first and is larger exactly when the ranking of articles changes between markets. That is what flexibility is, and where the ranking does not change there is nothing to buy. What the never-win column is for. A flexible machine is sold on the length of the list, and the list is not the case. An article that is never the best in any market this plant expects contributes nothing to the value of the option - it may still be worth making for a customer relationship or to fill a gap, but it does not belong in the argument for the machine. On the seeded plant, four articles are offered and two carry the entire case. Ask for the two, and ask what the machine costs without the other two. Why switching is charged the way it is. Each period is treated as an independent draw, so the plant is assumed to change its mind whenever a different article happens to win. Real markets are persistent - a bad quarter is usually followed by another one - which means fewer changes than this and a larger net value. The sheet errs against flexibility on purpose, because it is arguing for the expensive machine and should be made to work for it. What is not modelled. Four market states are a coarse picture of a continuous thing, and the answer is only as good as the four margins entered against them - a plant that cannot say what elastic tape earns in a bad market is not ready to price this decision. Nothing here values learning: a plant that runs two articles gets better at both and knows more about which way the market is going, and that is real and uncounted. Capacity is single-machine, so a plant that could buy two dedicated machines instead of one flexible one is asking a different question. And the option to make nothing at all - to stop the machine in a bad market rather than run it at a poor margin - is not on this sheet, and where a plant genuinely can idle a machine it is worth more than anything here.