More output from the same robots, with 40 people short

The plant

Two plants, more than 350 people and over 75 variants. Every finished tool is two matched halves that have to meet at final assembly, so a shortage of either half stops the whole order. Heat treatment runs on a robotic laser cell; forging runs on hammers with long changeovers.

The problem

• Output at heat treatment swung widely from day to day, and up to 60% of parts had to be run again.
• Work in process piled up everywhere, so the floor couldn’t see which matched pairs were really short.
• Each department was measured on its own output, which kept every machine busy on the wrong work.
• The plant was more than 40 people short through the year.

What we did

1. Found the bottleneck by loading, not by opinion. The numbers pointed to the heat-treat laser cell for the main product stream.
2. Changed the measures first. Departmental targets were replaced with one shipping-based productivity index, so every area pulled in the same direction.
3. Choked the release. Excess work in process was frozen and physically moved to a holding area. Raw material was released only for items that would run short.
4. Protected assembly with a buffer. Completed matched pairs were held at one stocking point in front of final assembly.
5. One priority list, every day. A five-colour priority, recalculated daily from a matched-pair shortage rule, owned by one named person.
6. Fixed the bottleneck’s losses. Failure-mode reviews and preventive measures cut the risk on the cell’s top failure modes (risk scores from 288 to 12 and from 140 to 30).

The results

• Retry rate at the bottleneck: from 30–60% to under 2%.
• Daily output through the bottleneck: from 6,000–25,000 to 30,000–35,000 units, on four of the five robots.
• Forge output: consistently over 150,000 pairs a week, up 25%.
• All of it achieved while the plant was more than 40 people short.

Another case study: The best shipping month in company history, three months in. See all case studies.