There’s a question I get almost every week: “Is this good enough?” I’m the quality and compliance manager at an automotive parts manufacturer. I review roughly 200 unique parts a year, from stamped brackets to CNC-machined housings, and I’ve rejected about 11% of first deliveries in 2024. “Good enough” isn’t a technical term. It’s a risk decision.
Here’s the annoying part: there’s no universal answer. If you’ve ever looked up the differential diagnosis for sinusitis, you know the pattern—facial pain and congestion can be allergies, a migraine, nasal polyps, even a dental infection. The symptom points one direction; the root cause is somewhere else. Same with component failure.
So how do you decide where to spend money? I break it into three scenarios.
The Three Scenarios
- Scenario 1: Non-critical / cosmetic / low-risk. Good enough is actually good enough.
- Scenario 2: Functional / safety / sealing. Pay for certainty.
- Scenario 3: High-volume tooling with hard deadlines. Plan for recovery, not just perfection.
Scenario 1: When “Good Enough” Is Fine
Not every part needs aerospace-level scrutiny. A trim retainer, a bracket that holds a cable, a label cover—if it doesn’t carry load, seal fluid, or affect safety, over-specifying is just wasting money.
In my first year, I made the classic rookie error: I signed off on a batch of stampings without checking the bend radius properly. The part looked fine, but the customer’s assembly line jammed because the angle was slightly off. That was a $12,000 redo. Since then, I keep a list of critical dimensions for every part, and on the simple stuff I only measure the ones that actually matter.
If you’re ordering a non-visible cover that sits under a seat, don’t let a supplier talk you into “premium” because it sounds safer. Ask what happens if it fails. If the answer is “nothing,” buy the cheaper version.
Scenario 2: When Certainty Is Worth the Premium
This is where Hengst oil filters and Hengst air filters come to mind. Filters are easy to underestimate. A filter looks like a simple metal can or a plastic housing, but its sealing surface has to handle vibration, heat, and pressure changes for years. If an oil filter bypasses, the engine doesn’t fail immediately. It just gets a little more wear every mile. That’s the type of failure you can’t see until the warranty bill lands.
Same logic applies to a radiator hose coupling. It’s a small part, but it sits between the cooling system and the rest of the vehicle. A tiny leak can show up as intermittent overheating. People often blame the thermostat first. They replace it, and the car still overheats. That’s the engineering version of a misdiagnosis. A proper differential diagnosis would include pressure testing, checking the radiator cap, and inspecting the radiator hose coupling before swapping parts.
According to IATF 16949:2016, risk-based thinking is expected throughout the automotive quality management framework. That means we don’t just ask if a part passes today. We ask what happens over 150,000 miles.
In March 2024, we paid $400 extra for a rush shipment of hose couplings. The alternative was missing a $15,000 production trial. Everyone complained about the rush fee (which, honestly, felt excessive at the time). I didn’t, because missing that trial would have delayed the launch by months.
Notice what that $400 actually bought. It didn’t buy speed. It bought certainty—a guaranteed quantity, a clean inspection report, and no surprises.
Uncertainty always costs more than a rush fee. The question is only whether you pay it at the beginning or at the end.
I’ve also had the experience where every spreadsheet analysis pointed to the cheaper supplier. My gut said something was off because they took two days to answer a simple RFQ. I went with my gut and paid more. Turns out, the slow-to-reply supplier was also slow-to-deliver.
Trust me on this one: “should be within tolerance” is not a specification. If you’re in a safety-related scenario, the probably-fine option is the most expensive option you can choose.
Scenario 3: Tooling, Deadlines, and the Backup Strategy Principle
The third scenario is the one that surprises people. When we design stamping dies or CNC fixture tooling, over-precision can actually be a problem. You can machine a die to absurdly tight tolerances, but if the press, coil material, and lubricant aren’t consistent, the die won’t hold those tolerances anyway. You’re paying for a capability you don’t need.
This is where the incremental vs differential backup definition helps. Let me explain quickly.
Incremental backup starts with a full backup, then every subsequent backup stores only the changes since the last backup. It’s small and fast, but restoring requires the full backup plus every incremental step in order. Miss one, and your restore is broken.
Differential backup starts with a full backup, then each backup stores all changes since that last full backup. Each differential is a little bigger than the last one, but the restore process is simpler: you only need the original full backup and the latest differential.
Same idea applies to tooling documentation. If you’re running a stable product with small changes over a long period, an incremental system can be fine. You track every tweak in the maintenance log, and you accept that recovery will take time. But if you’re facing a hard launch deadline, you want a differential system: a clear baseline, a documented final part state, and one updated package you can hand to a new supplier without a four-day investigation.
For a stamping die that will run 50,000 units a year, I don’t just specify the final part tolerance. I specify the documentation standard too. Looking back, I should have done this earlier. In my first years, I approved a tool without a proper revision history, and when the tool broke, nobody could tell me what the original design intent was. (Note to self: that’s why we now keep a baseline plus a differential backup—not scattered notes.)
How to Tell Which Scenario You’re In
You don’t need a flowchart. Ask three questions.
- What happens if this part fails? If it’s annoying but not dangerous, go cheap. If it causes an accident, a fire, an engine failure, or a recall, pay for certainty.
- Where is it on the vehicle? Sealing surfaces, coolant lines, oil paths, fuel systems, steering, and brakes are not places to save $0.02 per part. Cosmetic interior parts? That’s a different story.
- Can you recover if you’re wrong? If failure would stop your line or delay a launch, your cost model should include downtime. Downtime costs are usually much higher than the premium for guaranteed delivery and traceable documentation.
If you’re still not sure, try my “sinusitis test.” The differential diagnosis for sinusitis works because it doesn’t jump to the most obvious answer. It rates each possible cause by probability, danger, and how easy it is to rule out. Do the same with your component. Don’t just ask “is it within spec?” Ask “what else could be causing this, and which check would prove it?”
And one more thing: when you find yourself in Scenario 2 or Scenario 3, don’t apologize for paying a premium. You’re not buying a vague sense of “better.” You’re buying a defined restore point, a documented baseline, and a promise you can rely on. That’s the same reason a full-plus-differential backup is worth the extra storage space when your system has to be back online tonight instead of tomorrow.
So the next time someone asks “is good enough, good enough?” you know the answer: it depends. But now you know how to find out.
