The email came through at 7:47 on a Tuesday morning. Three CAD attachments, a short note, and a subject line that said, 'Custom stamping and machining quote—small run.' I'm the quality/compliance manager on the stamping side at Hengst, and I review every job before it ships. That's roughly 200 unique items a year, so I've learned not to judge an order by its piece count.
Marcus runs a two-person restoration company in Montana. He makes parts for old Ford trucks and Jeeps—the kind of stuff you can't get from a dealership anymore. His request had three part numbers: a stamped steel cover for an AMC 20 differential, a mounting bracket for an F250 fuel pump, and a CNC-machined spacer for an alternator rebuild kit. Total quantity: 500 pieces.
The order that looked too small to matter
In automotive terms, 500 pieces is not a production run. It's a trial lot. At a high-volume stamping plant, a cover like that could be formed in seconds, and 500 of them would be a warm-up job. I get why some suppliers skip these inquiries. The setup time, the sharpening of the trim die, the dimensional report—all of that costs more per part than it would on a 500,000-piece order.
Early in my career, I probably would have let it sit in my inbox. I remember losing a customer to that exact mistake. He needed a quote for 300 brackets, and I was too busy chasing bigger opportunities to reply. When I finally called him, he had already given the job to another shop. It wasn't price. It was my indifference. That lesson stayed with me.
So when Marcus wrote, 'I know it's a small run, but I need it right,' I took him seriously.
The Hengst catalog only covered part of the request
Marcus said he found us through the Hengst catalog. The F250 fuel pump bracket already existed as a standard blank from an earlier program, and he wanted a revised version with an extra mounting hole. That part made sense to our engineers. They reworked the CAD model, added a formed offset, and the tooling path was straightforward.
The AMC 20 differential cover was not in the catalog. His drawing specified a thicker flange and a gasket surface that had to stay flat enough to seal. That required a new draw die and an extra milling pass after stamping. Not a huge challenge, but not a standard catalog item either.
The third piece was the one I almost missed: a spacer for an alternator rebuild. Marcus wanted the kit ready before he posted one of his 'how to rebuild alternator' videos. The spacer had to locate the rear bearing in the housing. If the bore drifted by even a few tenths of a millimeter, the shaft would bind once the alternator got hot. The drawing said 'press fit,' but it didn't include a geometric tolerance for roundness or concentricity. That's the kind of detail that usually causes problems after the parts are already in customers' hands.
Then the first samples made it awkward
We ran the first samples in March. The F250 fuel pump bracket passed its dimensional check without much fuss. The alternator spacer measured within spec on the bench. But the AMC 20 differential cover had a trim burr along one bolt hole, and the mounting face showed 0.3 mm of waviness. It was inside the print tolerance, but it wasn't inside what I wanted to ship.
The production manager said I should release it because it met the drawing. The sales lead said Marcus was willing to accept it as-is. Marcus himself sent a text: 'We can deburr these at assembly, no big deal.' On an aftermarket restoration part, that would probably be true.
But his first email kept nagging at me. He said he wanted 'Hengst oil filter quality' on a piece of stamped steel. Most people know Hengst for oil filtration, but we also build stamping dies and machined components. When he said that, he wasn't talking about cosmetic finish. He was talking about consistency. It's tempting to think a small order can follow a lower standard, but that ignores the part where the customer is going to compare us to every other supplier he's ever used.
Why I stopped a shipment the customer already approved
I put a hold on the cover and pulled the tooling team in. The waviness came from inconsistent draw pad pressure. The burr came from a trim die that had been set up for slightly thicker material. Both issues are common. They're also easy to miss if the inspection report only checks the main dimensions.
I also knew our own internal spec was stricter than the customer's print. Under our IATF 16949 quality system, we use a 0.15 mm flatness limit for sealing surfaces, even when the drawing allows 0.25 mm. That boundary came from a Q1 audit where we found that parts near the high end of the tolerance sealed fine on a workbench but leaked intermittently after heat cycling. The drawing said acceptable. My experience said otherwise.
I didn't ask Marcus to accept it. I told him we were going to rework it. To be fair, he didn't argue. But I knew the delay hurt his schedule, and I knew it was my call to make.
What the rework cost—and what it bought
We reworked the trim die and adjusted the draw pad setup. The replacement AMC 20 differential covers came off the line at 0.08 mm flatness, with clean edges and no visible burrs. Marcus checked one with a straightedge, sent a photo, and approved the shipment.
The whole rework added about two weeks and $4,200 in labor and material to a job that had already eaten its margin. On a 500-piece order, that stings. I remember sitting in the cost review and explaining why I rejected parts that the customer had said he would accept. The only argument I had was that I didn't want our name on something I wouldn't put on my own truck.
But here's the part nobody can measure on a P&L: that small customer didn't disappear. In 2025, Marcus's total orders at Hengst were four times the value of that first purchase order. The AMC 20 differential cover is now a standard product on his website. The F250 fuel pump bracket has become a catalog option. And the alternator spacer led to a second variation for a different vehicle family.
The small-customer lesson I keep coming back to
Small doesn't mean unimportant. It usually means early. I'm not saying every 500-piece order needs the same paperwork as a full PPAP Level 3 submission. That would be overkill, and I've seen quality departments hide behind procedures instead of making decisions. But every order, no matter the size, has to clear the same basic question: would I send this out of my own shop?
I still keep one of the rejected covers on a shelf, burr and all. The reason isn't to shame anyone. It's to remind myself that making a good part is easy when everything goes according to plan. The harder part is explaining to a customer that 'good enough' isn't good enough, especially when he's already said it is. That's the part of quality inspection that never shows up in a control plan.
