Takt. That odd, German sounding word that often gets misspelled as tack and raises eyebrows from folks who haven’t breathed in a lifetime of production fumes. But to true manufacturing types, takt time is a vital concept for setting up production floors, modelling costs, and making continuous improvements.
Takt actually is an old German word that means flow or pace, and was traditionally used in the sense of a conductors baton. But in the twentieth century, driven both by prewar German manufacturers and postwar Japanese manufacturers, notably Toyota, takt time took on a different meaning
In production, Takt time is defined as the time between production units coming off a line. So if the takt time for widget manufacturing is 60 seconds, this means that every minute a unit is produced. The math can easily be inverted with production rate being defined as 1/(takt_time). So a takt time of 60 seconds means a production rate of 60 units / hour.
The reason takt time is important is that in doing line set up we typically work backwards. If we need to build 600 parts per eight hour shift, takt time is easy to calculate:
- Takt time =( 8 hours * 3600 second / hr )/ 600 = 48 seconds
- (Note that this neglects breaks, down time, efficiency, etc which all really need to be considered in a more exact calculation)
This means we need to produce one unit every 48 seconds.
Now suppose we know that it takes 8 minutes of labor (= 480 seconds) to build each widget. But there are many ways to accomplish this:
- Create 10 work cells, each with one operator, each taking on 1/10th of the build effort. This is the classic assembly line method
- Have 10 workers each build a unit end to end in parallel. Every 480 seconds we get 10 units, which yields the requisite 48 second takt time. This also works
- We can also do hybrid models, with say 8 workstations with one operator and one with two working in parallel
- There are other levers we could pull, including going to two shifts of 5 operators each
- Workcells can do varied tasks — assembly, test, pack out etc
- One operator could run two physical work stations and / or multitask (do something different during an unmanned test cycle for example)
- Labor content does not always equal production time. You can have cure times or unmanned test times or run in or whatever that have elapsed time without labor. This still, if set up properly, can yield the proper takt time. Suppose we have a 24 hour burn-in test in the above example. We still will get one unit off the line on average every 48 seconds, but only after a 1 day lag…
- Automated assembly also uses takt time–operators do not have to be humans (or mountain trolls or wood elves… just saying)
- The devil in either manned or unmanned cases are in the details, and often times it takes very careful drafting of work cells and flows to avoid inefficiencies where one operator takes only 40 seconds to do her / his task and then sits idle for 8 seconds…
- Takt time also drives the labor quantity, again be it machine or human or wood elf. (No of course we don’t hire wood elves at Zebulon Solutions, that would be illegal. But then again they do work for dandelion stems…)