Measuring performance in industrial component makers: what to track and what to ignore

Ask a component unit owner how a setter is performing and the answer often comes as a feeling. He is good, the machines run when he is on shift. Ask about an inspector and it is vaguer still. Feelings are not useless, but they cannot be compared across shifts, defended in a grade review, or used to spot a problem before a customer does.

The opposite mistake is common too. A plant buys a dashboard and measures everything it can count. Pieces per hour, machine running time, attendance, overtime, training hours. The supervisor is then buried in figures that mostly describe the machine, the order mix or the calendar, not the person.

Three tests sort a useful measure from noise. The person can influence it through their own work. It can be read for a machine and a shift, not just the whole plant. And it links to something the customer or the owner cares about: good parts, on time, without sorting.

• First-off cleared without a reset, per setter. After a changeover on a header, hobber, coiler or CNC lathe, did the quality room pass the opening parts straight away?
• Changeover time against an agreed band. Measured from last good part of the old job to first approved part of the new one.
• Internal rejection and rework by machine and shift. Not plant totals. Read at the machine, the figure tells you where to coach.
• Escapes by inspection stage. When a customer finds a fault, which stage did the lot pass through, and what did that check look at?
• Sorting hours. Time spent sorting suspect lots is a direct cost of upstream misses.
• Skill ladder progress. How many machines or tests each person is signed off on, and how many are in training.
• Night shift rejection compared with day. A gap points to cover and supervision rather than to individuals.

• Pieces per hour compared across machines. A multi-station header and an old single-die machine cannot be compared, and the order mix changes daily.
• Hours on the floor as effort. Long hours often mean poor planning or rework, not commitment.
• Annual star ratings from the supervisor. One number a year, set from memory, mostly measures who the supervisor likes.
• Suggestion counts. Counting suggestions without tracking which were acted on teaches people to write rather than think.
• Training hours. Classroom hours prove nothing about running the machine unsupervised.

The track list holds across the group, but the evidence behind each line changes with the part. Supervisors should know which record proves it in their own section.

• Bearing rings. Ovality and taper after grinding, superfinish marks, and the noise or vibration test result per batch.
• Gears. Tooth charts from the gear tester after each hob or shaving cutter change, and dimension over pins.
• Springs. Free length and rate on the load tester, and presetting losses noted per coiler.
• Hydraulic cylinders and blocks. Bore taper after honing, particle count at the wash station, and holding pressure on the test rig.
• Pipe fittings. Flash and lap marks after forging, socket bore size, and the hydrostatic test log.
• Fasteners. Head cracks found at the heading stage, thread gauge results after rolling, and hardness after the furnace.

Section and month: ____.
Setters: first-off approved at first attempt, by name: ____.
Changeovers outside the agreed band, with reason: ____.
Rejection and rework by machine, day and night: ____.
Customer escapes and the inspection stage each passed: ____.
Sorting hours spent on suspect lots: ____.
Sign-offs gained on the skill ladder this month: ____.
People who need coaching, and on what: ____.
People ready for a grade step: ____.
Actions agreed and owner: ____.

1. 1

   Agree the measures with supervisors

   Start with four or five from the track list. Explain why each one matters and what the supervisor will do with it.

2. 2

   Set up the data at the machine

   A rejection tag that records machine, shift and operator; a changeover log beside the setter's bench. The pneumatic parts reporting page shows a light way of collecting this.

3. 3

   Review monthly, one supervisor at a time

   Thirty minutes with the sheet. It should end with coaching actions and any grade steps, written down.

4. 4

   Link to the grade ladder, not to blame

   Use the sheet to decide who can move up to another machine or to setter. The shaft shop review page and the pipe fitting equivalent show trade versions.

5. 5

   Hold a short twice-yearly conversation

   Each person sees their own figures, hears what has improved, and names one skill to add before the next talk.

Posting an operator league table is the classic misstep. People at the bottom stop reporting their own rejects, and the numbers get worse while looking better. Show figures by machine and shift on the floor, and discuss individuals privately. The second failure is changing measures every few months. Pick a set, live with it for a year, and adjust only when a measure is clearly misleading.

Good measures make supervisors better coaches. They also make a grade decision easy to explain at the bench. Our appraisal and review cycle design can fit the sheet to your sections. The hub for bearing, gear, valve and fastener makers sets the context.