Skip to content
Improvement Kata and Coaching Kata | Part 2 of 6

Grasping the Current Condition: the Spread Carries the Information

How to measure how a process works today, how many cycles it takes, and why the widest step matters more than the slowest one.

Ask a supervisor how the line runs and you get an opinion with a number in it. Ask for the record and you usually get a monthly report of what was booked. Neither is a current condition. A current condition is how the process works today, measured at the process, in enough detail that the next target condition has something to break.

What to measure, and how much

The result measure is the figure the challenge is written in. On L3 that is units per shift, taken from ten shifts rather than from a good day. The process measure is what the experiments will actually move: seconds per cycle and station, from at least sixty cycles across variants and shifts. Then the spread, the pattern sorted by station, variant, shift and hour, and the interruptions with the place where they start, because every stop is a cycle that never happened.

What L3 looked like in week one

Station Range in seconds Inside 75 s Stops per shift Mean seconds
01 Frame loading 58 to 81 87 % 1 64
02 Welding fixture 60 to 104 72 % 3 71
03 Hood and seals 58 to 112 73 % 1 69
04 Cable routing and test 52 to 141 38 % 14 84
05 Panel mounting 59 to 92 82 % 0 66
06 Labeling and documents 48 to 96 88 % 0 58
07 Final test 61 to 118 68 % 4 73
08 Pack out 52 to 85 90 % 0 61

The line produced 310 units per shift, an effective cycle of 87 seconds against a planned 75. The 23 stops cost 960 seconds of the shift, about 40 seconds each.

What the numbers said

Station 4 is not the slowest station on average. Three stations sit within ten seconds of it. Station 4 is the widest: 52 to 141 seconds, 38 percent of cycles inside the planned time, and 14 of the 23 stops per shift start there. The five cycles above 120 seconds were all the same variant.

That is the finding an average would have hidden. A mean of 84 seconds looks like a balancing problem and invites a discussion about moving work to the neighbor. The spread says something else: the same person does the same job in 52 seconds and in 141, so the problem is not the work content.

The second finding is quieter. With every station inside 75 seconds, 360 units fit into the shift. The line is not short of capacity. It is short of stability.

Three ways to get it wrong

Collecting the average only. The improvement lives in the tail, and the tail is where the stops come from.

Measuring from the system instead of the process. The system reports what was booked at the end of the shift. The cycle is what the operator does, and the two differ by everything that is interesting.

Measuring once and never again. A current condition older than the last target condition is a memory, so the process is measured again before the next one is set.

Write it on one sheet

The current condition belongs on one sheet at the process, by hand: the result measure over ten shifts, the process measure per station with lowest, highest and share inside the planned cycle, the pattern in one sentence, and the count of interruptions. On L3 that sheet took two people half a shift to fill and stayed on the record board for all sixteen weeks. It is also what makes the next step possible, because a target condition can only be written against a condition somebody has measured.

Newsletter

One Lean idea.
Every two weeks.