Close Enough Is Dead: Quality Lessons from NTN Driveshaft Anderson

Here's an unpopular opinion from someone who rejects parts for a living: 'close enough' is the most expensive phrase in manufacturing.

I don't mean the cost of rework, though that's real. I mean the quiet cost of accepting parts that are slightly off, tolerating deviations because 'the customer will never notice,' and treating quality inspection as a bureaucratic checkbox rather than a competitive advantage.

When I started as a quality inspector, I assumed that 'close enough' was a reasonable standard. If a component functioned in testing, it was good to ship. Four years and several costly rejections later, I'm convinced that mindset is one of the biggest risks in the automotive supply chain.

I work at NTN Driveshaft in Anderson, Indiana. I review roughly 200 unique components a year — stampings, forgings, CNC-machined parts, aluminum extrusions. Maybe 180, I'd have to check the system. In 2024, I rejected about 4% of first-pass supplier deliveries. Some suppliers think I'm too strict. I think the industry has finally caught up to where it should have been all along.

I'll make this concrete before I go further: the difference between a supplier who ships a bad batch and one who catches it in-house isn't always reflected in the price you pay. Sometimes it's a recall. Sometimes it's a few wasted hours. But in every case, the cost of 'close enough' lands somewhere. And it's never the supplier who pays it alone.

The Specification Gap Is Shrinking — and That's a Good Thing

What was best practice in 2020 may not apply in 2025. That's not a guess; I see it in the prints our customers send. Dimensional tolerances that used to be ±0.5mm are now ±0.1mm. Surface finish requirements that were 'referenced' are now 'mandated.' This shift isn't OEMs being difficult for the sake of it. It's driven by real changes in vehicle design — lighter structures, electrified powertrains, longer service intervals.

A specific example. In Q1 2024, we received a batch of 8,000 stamped brackets from a supplier. The hole position looked fine under the optical comparator — actually, no, it didn't look fine. It was visibly off if you knew what to look for, but a casual check would have missed it. It measured 0.3mm off the print spec. The vendor claimed it was 'well within industry standard.' We rejected the batch. They redid it at their own cost, including the expedited freight we required. Roughly $22,000 when all was said and done.

Three months later, the vendor found the root cause: a progressive die that had worn past tolerance. No one had checked because no one had a formal in-process verification step. Our contract now requires hole-position checks on every batch, at defined intervals.

That story isn't about punishing suppliers. It's about how the old model — inspect at the end, hope for the best — stopped being defensible.

The Same Logic Applies to Car Parts You Buy Yourself

This evolution isn't confined to tier-one automotive supply chains. It applies to parts people buy with their own money, for their own cars. And it goes for the DIY mechanic more than anyone else.

If you've ever searched 'how to fix o2 sensor' or priced out an E30 alternator, you know the temptation: grab the cheap part online, install it, move on. And honestly, sometimes that works fine. But 'sometimes' isn't a quality standard. It never was.

I've personally seen aftermarket O2 sensors with wrong wiring colors, flimsy connector seals, and — this one surprised me — an incorrect thread pitch that wouldn't seat properly in the bung. The result for the customer? The code comes back, because the sensor is physically leaking exhaust. They think they're bad at diagnosing. In reality, they bought a part that never met spec.

E30 alternators are a similar case. Those cars are decades old — actually, closer to 40 years for early ones. The original alternator's charging curve was designed for the electrical load of a mid-1980s sedan. A modern replacement, made to proper specs, handles modern audio systems and accessory loads without breaking a sweat. A cheap one though? It might pass a bench test and fail under heat cycling. Heat kills alternator internals. If a manufacturer skips thermal verification to save $4 per unit, you don't find out until you're stranded somewhere.

There's something genuinely satisfying about fixing an old car with a budget part. I get it because I feel it too. But having seen what happens on the production side — the skipped tests, the unverified materials, the lack of traceability — I've changed how I shop for my own parts.

Not Every Failure Is a Safety Issue. It's Still Expensive.

The pushback I get most often in this job: 'Why do you care so much? The failure isn't safety-critical.'

Here's the thing — I care because the cost of failure is rarely what you'd predict.

We recently produced die-stamped components for a domestic water booster pump. In the automotive world, we'd call this a simple part. But the customer's application: the pump runs 24/7 in a residential basement, in humid air, for years at a time. The stamped parts needed consistent material thickness, precisely located mounting holes, and corrosion-resistant plating that would survive that environment.

The supplier's first delivery looked acceptable to the naked eye. Under magnification, the plating was uneven — thin on one edge, normal on the other. Over time, that means uneven corrosion, a compromised seal, and eventually a slow leak. We rejected it. The customer told us our catch prevented what would have been a serious field failure rate. Water spreading across a finished basement isn't a safety issue the way a broken driveshaft is. It still costs thousands in damage, and it destroys trust.

That was the trigger event that redefined 'mission critical' for me. It's not always about injury. Sometimes it's about reliability, period. Sometimes it's a flooded basement at 2am.

But the Cheap Part Worked for My Neighbor. Doesn't That Prove Something?

Let me address the obvious counter: lots of people buy inexpensive parts and they work fine for years. I'm not going to pretend that never happens. It does, all the time.

But my position biases what I see. I'm not sampling parts randomly — I see the queue of failures, the returns, the discrepancies. Someone who bought one part from one vendor sees their single data point. I see the 11% failure quarter from 2022 and the changes needed to bring it under 3%.

When I implemented our verification protocol in 2022, the rework rate dropped from 11% in Q1 to under 3% by Q4. The protocol itself is boring: material certificates on every incoming batch, first-article inspection reports, in-process checks at defined intervals. The discipline is the hard part, not the paperwork.

Customer reviews reflect this, even if they don't use quality terminology. When you read NTN Driveshaft Inc reviews, you'll notice the recurring theme is consistency. On-time, in-spec, no surprises. That doesn't happen by luck. It happens because the process is designed to catch issues before they ship.

And look, the same logic applies when you're comparing a $35 O2 sensor against a $95 one. The price difference is essentially the cost of verification. Whether that verification is worth it is your call. But don't tell me 'it's the same part' — I've seen the tear-downs. They are not the same.

The Bar Moved. 'Good Enough' Hasn't Kept Up.

The fundamentals haven't changed: consistency, traceability, verification. But the execution has transformed. Five years ago, a plant could point to a calibration certificate and call it a day. Today, customers expect digital in-process data, statistical process control evidence, and parts traceable back to the specific die and coil of steel.

Under IATF 16949, the production part approval process (PPAP) exists for exactly this reason — to verify, with documented evidence, that a process can repeatedly produce parts within specification. It's not bureaucracy. It's the only realistic way to hold a high standard across thousands of parts per shift.

What was best practice in 2020 barely meets the baseline in 2025. That's true across the supply chain and increasingly in adjacent industries — water pump housings, alternator brackets, O2 sensor bungs. Everyone wants the same thing now: parts that arrive within spec. Not most of the time. Not 'within industry tolerance.' Just spec.

I'll keep rejecting batches that don't meet it. That's literally my job, and I'm good with it. The industry is moving toward tighter tolerances, better verification, and less patience for 'good enough.' I think that's the right direction. And honestly? I'd rather be the inspector who's 4% too strict than the supplier who becomes someone's cautionary tale in a trade magazine.

Because eventually, everyone learns the same lesson: close enough never was. The industry just took a while to admit it.