SMED — Single-Minute Exchange of Die — reduces the time a process is stopped while switching between products or configurations. The name refers to the original target of getting changeovers under ten minutes.
The method rests on one distinction. Internal work can only happen while the process is stopped; external work could be done while it is still running — fetching tooling, preparing materials, pre-setting. Observe a changeover, classify every element as one or the other, move as much as possible to external, then streamline what genuinely remains. Most first-pass reductions come from work that never needed the stoppage at all.
The reason to care is not the minutes. Long changeovers force large batches, because the setup cost has to be spread across more units, and large batches mean more inventory, longer lead times and later defect detection. Cutting changeover time makes small batches economically viable, which changes the behaviour of everything downstream.
Which is why optimising a changeover performed twice a year is a solved problem that did not need solving, and why the target time is not the goal. Flexibility is.
