Measuring performance in sheet metal and fabrication units: what to track and what to ignore

An owner starts ranking his welders by kilos of MIG wire drawn from stores each month. Top of the list is a quick, cheerful hand. He is also the one laying fillets twice the drawn leg, leaving a carpet of spatter for the grinder and re-welding his own wormholes every other week. At the bottom sits a quiet welder whose correctly sized joints never return. The ranking crowned the wrong man.

A fabrication shed throws off tallies all day: spools issued, sheets cut, tonnes loaded, strokes counted. Most describe activity. A handful describe what matters: whether parts fit at the next station, whether jobs leave on the promised day, whether skill is growing. Picking that handful sits at the centre of how we run a performance management process. It grows straight out of the rework pattern on the hub for fabrication HR.

• Clean first passes, by bay: jobs clearing your final check with no correction slip. Taken from slips and the gate-out file.
• Correction hours caused, by station: from the slips, filed under the station where the fault started, not where someone spotted it.
• Gate-out against promise: each job's promised date beside the date on its delivery challan.
• First-offs signed before bundles ran: signed and unsigned first-offs at every changeover.
• Changeover time on the bottleneck: press brake tooling swaps or stamping die changes, clocked by the operator himself.
• Beam-on hours for lasers: cutting time against staffed time, read off the controller and discussed carefully.
• Grade tests passed: retests sat and cleared each quarter.
• Near misses reported: read as a sign of trust, not of danger.
• Snag items at erection or installation: for site jobs, the client engineer's snag sheet per job.

• Kilos of wire or electrode packets per welder. Heavy use can mean oversized fillets, spatter and repairs.
• Hours on the premises. Someone staying late redoing his own fit-up looks devoted and costs you twice.
• Joints, brackets or panels counted alone. Without acceptance, the count pays for haste.
• Raw tonnes loaded. It hides whatever the erection gang sent back.
• Strokes per shift on a power press. Handy for planning, hazardous as a personal target.
• Sheets cut per shift. Silent on nest density, remnant use and wrong cartons.
• Loose ratings for attitude. Swap them for observed behaviour: checks done, help given, guards respected.

Most fabrication is flow work. A blank travels from shear to brake to fit-up table to welding bay to grinding corner, and faults ride along. Judge flow work by bay and station, then let the slips show where a fault started. Keep personal readings for roles that work alone: the nester's sheet yield and remnant rack, the die setter's stopwatch, the polisher's panels accepted against the reference sample. Even then, talk the reading through with him before concluding anything.

• Shear and punching: blanks re-sheared because the cut list or the back stop was wrong.
• Press brake: angle drift along a bundle, read on the first, middle and last piece.
• Fit-up table: frames failing a corner-to-corner check before any welding.
• Galvanising run: jobs returned from the galvaniser for blocked vent holes, trapped flux or zinc runs.
• Packing and dispatch: cartons opened by the buyer and found short or mixed.
• Erection gang: snags still open on the day the client engineer walks the frame.

Every one of these points to a habit someone can change. A cut list checked twice. A protractor used mid-bundle. A tape stretched across diagonals. Vent holes drilled before dispatch. A packing list ticked line by line. A snag walk the evening before handover.

1. 1

   Pick a short list per bay

   From the watch list, take the four or five that bite there. A welding bay might use clean first passes, correction hours caused, gate-out against promise and grade tests passed.

2. 2

   Name one source for each

   Correction slips, job travellers, first-off sheets, the gate-out file, the laser controller. Any reading that needs fresh paperwork nobody will keep, drop.

3. 3

   Hang a board at the bay

   Chalk it up every Friday. Anyone in the crew should read it in under a minute.

4. 4

   Gather the crew weekly

   Ten minutes standing at the board: what went right, the biggest correction cause, one change to try.

5. 5

   Sit with each person monthly

   A brief one-to-one with the chargehand: his share of the board, progress toward the next grade, one thing to attempt, one thing the shop owes him.

6. 6

   Revisit grades quarterly

   Use tests cleared and monthly notes to decide who sits a retest. Pay follows grades, never one month's tallies.

Bay and week: ________.
Jobs through the gate, and how many on the promised day: ________.
Clean first passes, and jobs carrying a correction slip: ________.
Correction hours caused here: ________.
Biggest correction cause this week: ________.
First-offs signed before the bundle, and any missed: ________.
Tooling or die changeover time on the main machine: ________.
Near misses reported, and what got fixed: ________.
Grade tests sat and cleared: ________.
Next week's single change, and who leads it: ________.

The board belongs to the crew; the monthly sit-down belongs to the person. Bring the board, his slips and his grade record, and ask before telling. 'Corrections from the fit-up table jumped this week. What happened there?' tends to uncover a missing jig or a rushed order, not merely a careless hand. Two business types have their own angle. One is judging a welder's month beyond the bead count. The other is beam time against accuracy in laser job shops. Rolling boards into a monthly people report is covered in people reporting for general fabricators.

Those same correction slips feed bonuses, as quality-weighted incentives for fabrication explains. The slip routine itself appears in the post on finish and rework. Keep the list short and the sources honest, and a Friday board will reveal far more than any league table of wire spools.