Takt time

Takt comes from German and refers to the musical bar, the beat. Takt time is not a performance of the workshop: it is an external constraint, dictated by the customer. It gives the pace at which one would have to produce to deliver neither early nor late.

Three notions are constantly confused. Takt time is the pace demanded. Cycle time is the pace actually achieved at a station. Lead time is the time a part takes to go through the whole process. Comparing them is the essence of line balancing: a station whose cycle time exceeds the takt is a bottleneck, a station well below it produces waiting or overproduction.

Takt is used to size: how many stations, what content per station, what capacity to plan. It is a tool for designing an organisation, and it is read with a balancing chart that compares each station with that reference.

It is not used to drive people. Displaying a takt of 47 seconds above a station and asking for it to be held is converting a customer datum into individual pressure — and obtaining, at best, shortcuts on quality. When a station does not hold the takt, the question is about the content of the station, not about the operator.

Faced with a demand that varies, the answer is not to recalculate the takt continuously but to level: volume and mix are spread over the period so as to present the workshop with a regular load. That is the role of heijunka, and it is what makes takt usable in real life.

Assembly of electrical sub-assemblies, two shifts a day. Each shift opens 480 minutes, from which 35 minutes of breaks and 5 minutes of team briefing are taken off: 440 minutes available per shift, i.e. 880 minutes per day, or 52,800 seconds. The customer asks for 1,100 parts per day.

Takt time = 52,800 / 1,100 = 48 seconds. One part must come out every 48 seconds to deliver neither early nor late. That figure says nothing about what the workshop can do; it says what the customer expects.

The line has six stations, with recorded cycle times of 44, 47, 58, 41, 39 and 46 seconds. Their sum, 275 seconds, is the total work content of one part. Related to the takt, it gives the number of stations needed: 275 / 48 = 5.7, so six. The line is correctly sized — the problem is elsewhere.

It is at station 3, at 58 seconds, ten above the takt. That station sets the real output: 52,800 / 58 = 910 parts per day, i.e. 190 fewer than demand. The other five stations may well be under the takt, they can do nothing about it: they wait, or they pile up in front of station 3.

Rebalancing requires no additional station. By transferring six seconds from station 3 to station 5 and five to station 4, one gets 44, 47, 47, 46, 45 and 46 seconds. All pass under the takt, the most loaded settles at 47 seconds, and the output becomes 52,800 / 47 = 1,123 parts per day — demand covered with the same people and the same equipment.

The utilisation rate — 275 seconds of work related to the 288 that six stations at takt make available — stands at 95 %. The line is tight: thirteen seconds of cumulative margin over six stations to absorb disruptions. That is the figure to look at before promising 1,100 parts every day, because it says how fragile the promise is.

One last reading, which often changes the conclusion: if demand rises to 1,250 parts, the takt falls to 42 seconds and no six-station split holds, since 275 / 42 = 6.5. A seventh station is needed, or work content must be removed. Takt is precisely for knowing that before committing to the customer.