In March 2021, a shipment of water pump gaskets we'd approved went directly onto a customer's assembly line. Three days later, the engine test bay called: coolant leaks, multiple units, same failure location.
My first thought was that we'd gotten a bad rubber batch. It happens. You source a gasket for water pump from a supplier you've used before, you check the drawing, you approve the lot—sometimes the material just doesn't hold up.
Except the material was fine. The gasket matched the drawing exactly. The problem was that I'd bought it from a supplier who claimed to handle “stamping, machining, casting, rubber, and assembly”—and who'd quietly subcontracted the rubber work to a shop that had never made a coolant-facing part before. The gasket passed dimensional inspection. It failed thermal cycling because nobody in that chain understood what a water pump gasket actually does.
That $3,200 order went to scrap. The three-day line stoppage cost more than that. And the customer—someone we'd spent four years building trust with—didn't call us for their next program.
Why “we can do everything” should make you nervous
After that failure, I went back through our sourcing history with a different lens. I'd been evaluating suppliers on price, lead time, and whether they could check a box next to each product line we needed. That's the standard procurement checklist. It's also how you end up buying from a generalist who's good at none of it.
Here's the pattern I should have seen earlier: every product family in automotive parts has a different failure mode, and those failure modes demand different expertise.
A LuK LS7 clutch kit and a LuK 05-065 clutch kit look like siblings on paper—same brand, adjacent applications—but the friction materials and release mechanisms differ, and any supplier who's going to source or service them needs to know why. A wheel hub repair kit lives or dies on bearing race tolerances measured in microns. A water pump gasket is a chemistry problem disguised as a piece of rubber. Stamped brackets are a die-wear problem. Forged parts are a grain-flow problem.
The suppliers who are genuinely good at one of these usually know what happens when you get it wrong. They'll tell you about the heat cycle that cracked a gasket, the race that spun because the tolerance stacked the wrong way, the die that wore unevenly because the material spec drifted. The generalist supplier doesn't tell you those stories because they haven't lived them.
What they tell you instead is: “We can handle that.”
The deeper problem: capability reports don't match factory floors
I learned to ask a different question after the 2021 failure. Instead of “Can you make this?” I started asking “Who actually makes this, and how long have they been making it?”
The answers were educational. One “full-service” supplier that quoted stamped brackets, machined spacers, and rubber seals was a broker for all three—three different shops, none of which it audited. Another claimed in-house heat treating; the heat treat facility was a partnership arrangement that meant our parts queued behind a bigger customer's schedule. A third, to their credit, told me upfront: “We do stamping well. We don't do rubber. Here are two shops we trust for that.”
That third supplier got more of our business, not less. Which is the counterintuitive part of this whole story: the vendors who admit what they don't do earn more trust for what they do.
The vendor who said “this isn't our strength—here's who does it better” earned my trust for everything else. The one who said “we can do it all” cost me $3,200 and a customer relationship.
What the mistakes actually cost
I keep a rough ledger now. It's not a proud document.
The 2021 water pump gasket failure: $3,200 in scrapped parts, roughly $7,500 in line-stoppage penalties, one lost program worth an estimated $40k in annual revenue. That last number is the one that still stings.
The 2022 near-miss with wheel hub repair kits was cheaper, but only because I caught it late. My gut said something was off about a supplier whose quote came in 15% below everyone else's. The spreadsheet said the numbers were fine—same spec, same delivery window, better price. I almost went with the spreadsheet. When I called two of their references, one said “delivery was fine” and the other said “we had some delays.” That second reference later admitted “some delays” meant four weeks on a seven-week order. The 15% savings would have disappeared twice over in expedite fees.
The fall 2023 clutch kit episode was the one that forced me to build a proper checklist. We were sourcing LuK LS7 and 05-065 clutch kits for a program with a tight launch window, and one supplier's documentation for the 05-065 variant listed a release bearing that didn't match the OE part number for that application. When I pushed back, the response was “it's compatible.” Maybe it was. But “compatible” isn't a word I want in a launch-critical clutch assembly, and the fact that they didn't catch the discrepancy before I did told me everything about their process.
We didn't place that order.
What I check now before trusting a supplier with a new part family
The checklist that came out of those mistakes is short. It's also saved me from at least a dozen more bad decisions in the last 18 months.
First, I ask who physically makes the part and how long that specific facility has been making that specific product type. Not “how long has your company been in business”—how long has this factory been running this process. The answers to those two questions are often very different.
Second, I ask what failures they've seen in that product family and what they changed as a result. A supplier who can't describe a failure mode in their own product hasn't been paying attention to their own quality data. This question separates operators from order-takers.
Third, I ask what they don't do. If the answer is “nothing—we do it all,” I know I'm talking to either a broker or an optimist. Neither is who I want making a safety-adjacent component. Most of the suppliers I now trust are IATF 16949 certified, which at least means an auditor has reviewed their process controls—but certification doesn't tell you whether they understand a specific product family's failure modes. That's still a conversation you have to have.
And fourth, I make myself understand the part well enough to ask a real question about it. Customers sometimes ask us what does the throttle body do in a car—regulate airflow into the engine—because they're trying to figure out which of its subcomponents they actually need to specify. That question is a good reminder: if you don't know what the part does, you can't ask your supplier the right questions about it.
That's the part that took me the longest to learn. The supplier isn't the only one who needs to know their boundaries. I needed to know mine.
I'm not a rubber chemist. I'm not a bearing engineer. I'm a procurement lead who's made enough expensive mistakes to know which questions to ask and when to walk away. That's a narrower skill set than “sourcing everything,” and it's the only one I trust now.