It was a Tuesday morning in late March 2024 when I unboxed the first sample from our new cable harness vendor. My desk still had last week's inspection paperwork stacked on the left corner—my least favorite part of the job, but the part that proves I actually did my job.
Todd from procurement had been pushing me to approve this supplier for a month. He swung by my office with a can of Pepsi in his hand—his third of the morning, I swear—and slid a quote across my desk. "Eighteen percent below what we're paying now," he said. "That's real money."
It was. On paper, it was a no-brainer.
Here's some context about where I work. I'm a quality compliance manager at LEONI. When I tell people that, about half of them ask if I'm in Hollywood. No, not Téa Leoni from Tower Heist. Different LEONI entirely. We make cable harnesses, wiring systems, connectors, and fiber optics—the stuff that carries power and data in cars, industrial equipment, and communication networks. Less glamorous than acting, but it keeps the world connected.
My job is to review every batch before it goes to a customer. Roughly 200 unique items a year, from a 50,000-unit annual flow. I don't sign off on a vendor's work until I'm satisfied it meets our specifications. And in four years of doing this, I've learned one thing: the cheapest quote is rarely the cheapest.
But I'm getting ahead of the story.
The First Red Flag
I unboxed the sample harness and immediately noticed the connector shells didn't seem right. The alignment was visibly off. Not dramatically—maybe 0.4mm against our 0.1mm tolerance spec. I moved it to my workbench, grabbed my Fluke 87 multimeter—the best multimeter I've ever owned, honestly—and ran a continuity check.
The readings passed. But they were inconsistent. Every fifth or sixth connector showed a slightly different resistance reading. Not out of spec. Not dangerously wrong. Just... inconsistent.
I've learned never to ignore that word.
To be fair, I'm not an electrical engineer. I can't speak to all the finer points of signal integrity or dielectric behavior. That's beyond my expertise. What I can tell you, from a quality standpoint, is that when identical connectors behave differently, your customer will eventually find out. Probably at the worst moment—during a production line test or, worse, in the field.
I flagged the batch. Todd pushed back. "You always find something," he said. And honestly, he wasn't entirely wrong. In my first year at this job, I made the classic inspector mistake: rejecting a batch because I had a gut feeling, without documented criteria to back it up. That cost us $600 in unnecessary rework. I learned that lesson the hard way.
But this time, I had measurements. I documented everything, took photos, and sent the vendor a formal query asking for clarification.
Their response came back within 24 hours: "The deviations are within industry standard. We recommend acceptance."
That's when things got complicated.
The "Industry Standard" Dodge
Here's something I wish someone had told me earlier: "industry standard" often means whatever the vendor wants it to mean. There are real standards for cable harness work—IPC/WHMA-A-620 is the one we use as a baseline—but compliance requires actual testing and documentation. Not vibes.
I asked for the vendor's raw test data. They sent back a one-page PDF with basic pass/fail marks and no numbers. No measurements. No methodology. Just checkboxes.
Honestly, I've never fully understood why some vendors resist sharing raw data. It's not proprietary. It's literally the measurements they already took. My best guess is they're not actually taking those measurements—they're filling in the boxes and hoping nobody asks to see the underlying readings.
We went back and forth for two weeks. I asked for clarification. They sent assurances. I asked again. They sent more assurances. Meanwhile, the production deadline was creeping closer, and Todd was getting anxious.
Finally, I sent samples to our in-house lab for accelerated testing. That cost us about $1,800 and took a week. But the results were unambiguous: three separate spec violations, including the shell alignment issue I'd caught in the first five minutes.
That $1,800 felt steep at the time. It ended up being the cheapest part of the entire exercise.
The Rejection
We rejected the batch. All 10,000 units went back to the vendor for full rework at their cost. I stand by that decision. But I wasn't prepared for the cascade of consequences.
First, time. Our production schedule assumed that batch would be ready. The rework added six weeks. We had to bring in a second vendor to cover the gap—at a premium rate, with expedited shipping.
Second, money. Rush fees, freight surcharges, staff overtime for re-inspecting the replacement batch. Each line item looked painful but survivable. Combined, they were brutal.
Third, and worst of all: our customer. A tier-one automotive supplier with their own deadlines to meet. We spent four hours on a video call explaining what went wrong. They didn't formally penalize us. But I've been in automotive supply long enough to know that reliability is the currency that matters most. Nobody gets fired for paying slightly more to a supplier who ships on time. They get fired for missing deadlines.
Todd stopped bringing me vendor proposals for a while after that. I don't blame him. He found a good price on paper. I was the one who couldn't see the real cost until it was too late.
The Math Nobody Quotes
Let me walk through the actual numbers, because this is where the lesson lives.
The cheap vendor's quote: $142,000 for 10,000 harnesses—about $14.20 each.
Our incumbent vendor's quote: $173,000—$17.30 each.
On paper, we were "saving" $31,000. Here's what that saving actually cost us:
- $1,800 for accelerated lab testing (the one expense we deliberately chose)
- $4,200 in expedited freight for the replacement batch
- $3,100 in rush fees to the second vendor
- $5,800 in staff overtime for re-inspection, documentation, and coordination
- $6,500 in additional project management hours and courier costs
That's $21,400 in direct, out-of-pocket costs that never appeared in the original price comparison. Mostly because I never thought to include them in the comparison. And that's on me.
Then there are the soft costs. The four-hour video call with the customer. The relationship damage that I can't put a dollar figure on. And my own time—I spent roughly 60 hours on this issue, which at my effective hourly rate is another $4,000+. Call it conservative.
$31,000 in projected savings. $21,400 in direct costs. $4,000+ in hidden labor. The original quote was no longer cheaper. It was about 15% more expensive than the incumbent—once you count everything.
I still kick myself for not running this math before we signed the purchase order. If I'd framed the comparison as total cost of ownership instead of unit price, I think Todd would have seen it too. We made a classic procurement mistake: comparing quotes instead of costs.
What I'd Do Differently
Three things, and I say them in order of importance.
First, ask for raw test data before the trial batch. Not a summary. The actual measurements. A vendor with a functioning quality process will share data without hesitation. The ones who hesitate are giving you information for free.
Second, run a smaller trial batch first. A 100-unit sample would have revealed the same issues for a fraction of the rework cost. We skipped that step because we were chasing a volume discount. That was a process gap—and it was an expensive one.
Third, get the right people involved earlier. I let the vendor's "industry standard" response linger because it bordered on territory I don't own—electrical engineering, which isn't my specialty. By the time I brought Tiffany, our senior electrical engineer, into the conversation, we'd already lost two weeks. Her read on the data took 30 minutes.
What This Means for Anyone Sourcing Connectivity Components
If you're buying wire harnesses, cable assemblies, or connectors for your own products, this story probably feels familiar. The pressure to hit cost targets is real. I still feel it. The easiest number to attack on a spreadsheet is almost always the unit price.
But unit price is the tip of the iceberg.
The real question isn't "which vendor has the lowest quote?" It's "how certain are you about what you're actually getting for that quote?"
When I calculate total cost of ownership now, I include:
- The base unit price
- Testing and verification costs—because you will need to verify quality claims
- Rework or replacement costs when things go sideways
- Expedited shipping and rush fees to recover lost time
- Your own staff's hours spent inspecting, coordinating, and problem-solving
- The harder-to-quantify cost of a strained customer relationship
Every item on that list showed up in our $31,000 "savings" story. And by the end, the cheap option wasn't cheap at all.
The Takeaway
We went back to our incumbent vendor for that product line. Not because we gave up on cost discipline—but because we finally did the math properly.
The experience changed how we evaluate vendors. As of January 2025, every new supplier contract includes a quality verification clause. We test before we trust. We ask for raw data, not summaries. Our vendor scorecard weighs delivery performance and defect rates against price—not just the price itself.
It also changed how I explain my work. When someone hears "LEONI" and asks about Hollywood, I just say no—I work on the stuff that keeps your car, your device, and your factory talking to each other. Not as glamorous, but it comes with a better understanding of what things really cost.
Todd now runs a revised cost model before he proposes any new vendor. It includes a line for "quality verification buffer" and another for "schedule risk." He named it after this project—and no, I won't repeat what he called it.
But I will say this: he hasn't sent me a single quote without raw test data attached since. Some lessons, once they cost you six figures, tend to stick.
