Lead time at plant level is a good objective and a poor steering measure for a team leader in pre-assembly. The team cannot see it, cannot influence it alone, and only learns the result weeks later.
A KPI tree solves this by breaking the objective down into measures each level can act on.
Building the tree
Start with the objective at the top: lead time from 18 to 13 days. Then ask what drives it:
- stock between processes in days,
- changeover time on the bottleneck presses,
- schedule adherence in pre-assembly,
- waiting time for released orders.
Each driver can be broken down once more. Stock in days depends on lot sizes and the number of part numbers in pull. Changeover time depends on internal and external setup steps.
Stop when a measure reaches a level where a team can influence it within a week or a month. A good tree for one objective has eight to fifteen measures on three levels, not fifty.
Leading and lagging
Lead time is a lagging measure: it shows the result. The number of part numbers in pull is a leading measure: it moves first and predicts the result. A useful tree contains both. Teams steer the leading measures. The leadership team checks that the lagging measure follows. If it does not, the tree is wrong and needs revision.
The bowling chart
Each measure gets a bowling chart: one row for the plan, one for the actual, twelve columns for the months, and a color in each actual cell.
| Measure | Jan | Feb | Mar | Apr | |
|---|---|---|---|---|---|
| Part numbers in pull | Plan | 10 | 25 | 45 | 70 |
| Actual | 12 | 22 | 38 |
Green when the actual meets the plan, red when it does not. There is no amber in many plants, on purpose: either the plan holds or a countermeasure is needed.
The plan row is not a straight line. It follows the milestones from catchball. If the supermarket opens in March, the plan rises after March.
One page per objective
Put the KPI tree and the bowling charts of one objective on one page. In the monthly review, the owner walks through the page from top to bottom: lagging result, leading measures, red cells, countermeasures.
Common mistakes
- Too many measures: the review becomes a reading exercise.
- Measures without a plan row: green and red become opinions.
- Changing the plan when it turns red: the chart then measures the ability to adjust plans.
The next part covers what happens in the monthly review when a cell turns red.