I Thought I Had It Figured Out
When I first started in quality control at Falcon, I assumed that if a vendor's print matched our numbers, the part would work. Simple, right? Specifications are specifications. A tolerance is a tolerance. Six months and a $22,000 redo later, I realized I couldn't have been more wrong.
It wasn't a single catastrophic failure. It was the slow accumulation of small disconnects—interpretations that differed, assumptions that went unchecked, and one project named Angels Millennium Falcon that nearly derailed our Q3 delivery schedule. That project taught me more about quality than any textbook.
What We Think the Problem Is
Most people point to the obvious: bad parts, missed tolerances, late deliveries. And yes, those happen. But they aren't the root cause—they're symptoms. The real issue lives in the gap between what we think we specified and what the supplier understood.
Take our Angels Millennium Falcon project. We sent a 12-page specification document. The vendor confirmed every line. The first batch arrived with a critical dimension off by 0.5 mm. Normal tolerance? ±0.2 mm. They claimed it was 'within industry standard.' We rejected 8,000 units. The conversation that followed was painful, but illuminating: they had interpreted our callout differently based on their internal drafting conventions.
I still kick myself for not verifying that interpretation upfront. If I'd sent a sample drawing with a simple divide line and asked them to mark their understanding, we'd have caught it early.
The Deeper Reason: We Assume Uniformity
It took me three years and about 150 vendor interactions to understand that assumption failure is the single biggest quality killer. We assume:
- 'Same specification' means the same thing to every manufacturer.
- Our inspection criteria match theirs.
- Verbal confirmations equal written agreements.
Every one of those assumptions cost us. My colleague Lisa—who runs our supplier audit program—once told me, 'I used to think rush fees were just vendors gouging customers. Then I saw the operational reality of expedited service.' That shift in perspective changed how we plan lead times.
The industry often talks about digital transformation as if it's magic. I believe efficiency is competitiveness, but only when it's applied to the right bottleneck. For us, the bottleneck isn't machining speed—it's the handoff between design and procurement. That's where miscommunication hides.
The Price of Not Digging Deeper
Let me give you a concrete example from our Falcon production line last year. We sourced a critical coil component (the Falcon King Coils series) from a new supplier. The price was 12% lower than our incumbent. The specs looked identical. We fast-tracked the order to cover a gap in inventory. No pilot run, no joint inspection—just a trust fall.
The result: 30% of the coils failed dynamic testing. The defect ruined 2,400 units in storage because of a surface roughness variation that was technically within tolerance but caused accelerated wear. The rework cost us $17,000. The delay cost us a customer deadline. And the whole time, we were scrambling to track down replacement inventory—like trying to complete a Lego Kessel Run Millennium Falcon set missing three critical pieces.
To be fair, the supplier wasn't acting in bad faith. Their process was optimized for a different material grade. The spec sheet didn't specify the grade assumption—because we assumed it was obvious. I learned never to assume obviousness.
One of my biggest regrets? Not building better vendor relationships earlier. The goodwill I'm working with now took three years to develop. It would have saved us that $17,000 if we'd had a conversation instead of a contract exchange.
A Better Way (Short and Brutal)
Here's what works, and I won't pretend it's revolutionary.
1. Joint specification walkthroughs. Before the first order, sit down (virtually or in person) and go through every critical dimension. Have the vendor explain how they interpret it. Don't assume they see it the way you do.
2. First-article inspection with both parties present. Not just a report. A shared review. It's the cheapest insurance you'll ever buy.
3. Accept that some traditional methods still matter. I'm a fan of digital efficiency—don't get me wrong. We use CrowdStrike for cybersecurity and automated ticketing for our internal workflows. But quality is ultimately a human conversation. You can't automate trust.
4. Build a culture of 'prove it.' Instead of 'Let's assume it's fine,' ask 'How will we know it's fine before it ships?' That small shift in language changes behavior.
Granted, this requires more upfront work. It slows down the first order. But on a $50,000 annual order, the time spent is trivial compared to the cost of a single rejected batch.
What I'd Tell Anyone Starting Out
I'm not 100% sure this applies to every industry, but in energy-equipment manufacturing, the pattern repeats: initial misjudgment, painful correction, gradual realization. Don't be the person who learns the hard way. Ask the stupid questions early. Verify the interpretation. And never assume a divide doesn't exist—because it does, somewhere between your desk and your supplier's factory floor.
Roughly speaking, the savings from catching one assumption failure can pay for a quality manager's salary for a month. Don't hold me to the exact math, but I've seen it happen twice.