- \n
- Capacity depends on changeover hours per week, not duration alone.<\/li>\n
- Smaller batch sizes raise changeover frequency and use up the gain.<\/li>\n
- Saved changeover time becomes output only on a bottleneck.<\/li>\n
- Decide in advance what the released time will be used for.<\/li>\n<\/ul>\n<\/div>\n
What swissLean writes about changeover reduction<\/h2>\n
In R\u00fcstzeiten reduzieren in der Produktion Schweiz: So geht es<\/a> (Reducing changeover times in Swiss production: how it works), swissLean explains how plants shorten their changeovers with SMED and closes its five steps with a clear principle: “Was nicht gemessen wird, wird nicht verbessert.” (“What is not measured is not improved.”)<\/p><\/blockquote>\n
The five steps in the article are practical: separate internal and external setup time, prepare in advance, standardize procedures, reduce movements and handling, and measure the improvement continuously. As a practical example, swissLean describes a manufacturer that cut its changeovers from 45 to 20 minutes without new machines. The article gives the method for the individual changeover. This article adds a question that almost always comes after the project: how often does the line change over, with what batch sizes, and how many extra hours does production really get afterwards?<\/p>\n
Changeover minutes saved are not yet capacity gained<\/h2>\n
After a successful SMED workshop, one number usually gets reported: the new duration per changeover. For machine capacity, another figure matters: changeover hours per week. That figure comes from two factors, the duration of each changeover and the number of changeovers. If you only measure the first factor, you cannot tell whether the machine produces more after the project.<\/p>\n
Three effects often shrink the gain on paper:<\/p>\n
- \n
- Changeover frequency moves.<\/strong> When changeovers get easier, planning changes over more often, deliberately or not. The minutes saved per changeover are then used up by additional changeovers.<\/li>\n
- The run-up is not counted.<\/strong> The machine is running again, but not yet at nominal rate, or the first parts are scrap. Even if the changeover is logged as finished, the time until stable production is still lost.<\/li>\n
- The machine is not a bottleneck.<\/strong> A machine that already waits for orders or material turns saved changeover time into waiting time, not into output.<\/li>\n<\/ul>\n
None of this questions SMED. It only shows that shortening the changeover and releasing capacity are two different results, and they have to be measured separately.<\/p>\n
![\"Four-step](\"https:\/\/teeptrak.com\/wp-content\/uploads\/2026\/11\/teeptrak-swisslean-en.png\")
<\/figure>\nThe math: changeover duration times changeover frequency<\/h2>\n
An example to illustrate: a machine changes over ten times a week and the changeover duration drops from 40 to 25 minutes. That is 150 minutes, or two and a half hours per week. A second machine with the same improvement changes over only twice a week and gains 30 minutes. Same method, same improvement per changeover, but a factor of five in time released. Changeover frequency is part of deciding where a changeover project pays off most.<\/p>\n
The table below shows typical situations and what a shorter changeover brings in each case. These are general patterns, not results from a specific plant.<\/p>\n
- The run-up is not counted.<\/strong> The machine is running again, but not yet at nominal rate, or the first parts are scrap. Even if the changeover is logged as finished, the time until stable production is still lost.<\/li>\n
- Changeover frequency moves.<\/strong> When changeovers get easier, planning changes over more often, deliberately or not. The minutes saved per changeover are then used up by additional changeovers.<\/li>\n