The Unit Price Trap

I've spent eight years managing material procurement for a 180-person contract manufacturer. Polycarbonate, TPU, polyurethane, the occasional PC/ABS blend. My budget runs about $2.4 million a year, and I've tracked every purchase order in the same spreadsheet since 2017.

Here's what that spreadsheet taught me: the lowest quote is usually the most expensive decision in the room.

Not always. But often enough that I stopped treating unit price as the primary number.

This isn't a theory I picked up at a conference. It's a conclusion I resisted for about three years.

Where the Money Actually Goes: Yield, Not Discount

Take TPU yarn. We supply coated industrial fabric to two customers, and the TPU is a meaningful line item. In 2021, we ran a split lot across three suppliers: two generic grades and Covestro's Desmopan line. The quotes spanned about 9% from top to bottom. On a 4,000 kg order, choosing the cheapest option would have saved us roughly $2,400.

We didn't. Here's why.

The problem wasn't the resin spec on paper. It was lot-to-lot variability in melt flow. The cheapest supplier's material ran fine on Monday and jammed the extruder on Thursday. Our scrap rate climbed from 2.1% to 5.4% over the quarter. Rework, machine downtime, expedited replacement — I ran the numbers twice because I didn't believe them. The 9% "savings" cost us roughly 14% more per finished meter.

When I compared the three suppliers side by side over a full quarter — same product, same line, different resin — I finally understood why our senior process engineer kept repeating the same line.

"Cheap resin is never cheap." — our senior process engineer, who was right and enjoyed saying so.

The Covestro grade wasn't the lowest quote. It was the one that held tolerance across the run. That's a different product attribute, and it doesn't show up on a datasheet.

Certification Costs More Than the Resin

Aviation work is where this gets painful.

We make interior brackets and small structural covers for a Tier 2 aerospace customer. The polycarbonate in those parts has to meet flammability and smoke-density requirements under FAR 25.853. That's not a paperwork formality. It's a qualification cycle — burn testing, OEM sign-off, first-article inspection, and a paper trail that runs for months.

In 2022, we tried to switch PC suppliers to save about $900 on a 1,200 kg annual buy. The new resin met the published datasheet properties. What it didn't have was our customer's existing approval. Requalification took five months and cost us around $6,400 in testing and engineering time — and we still ended up reverting to the original supplier because the new material's smoke performance was marginal at the top of the thickness tolerance.

They warned me about requalification. I didn't listen. I saw a $900 line-item reduction and treated the approval process as a formality. It wasn't.

Plastic card substrates work the same way, just with different acronyms. Card bodies typically need to meet ISO/IEC 7810 for physical characteristics and, if there's a chip, ISO/IEC 7816 for the interface. The material also has to survive lamination, hot-stamping, and whatever chemical bath the personalization line runs. I've seen a "compatible" PC film save $0.04 per card and then fail lamination adhesion on 30,000 units. That's not a discount. That's a recall.

Market Share Is a Supply-Chain Metric, Not a Bragging Right

People ask me why we use Covestro for a chunk of our PC volume when Teijin or SABIC can quote lower on certain grades. Fair question.

The honest answer: it's not about the logo. It's about capacity and continuity. A handful of producers — Covestro, SABIC, Teijin, and Mitsubishi among them — hold most of the global polycarbonate capacity. When you look at publicly reported market share rankings, Covestro and SABIC usually sit near the top by volume, with Teijin, Mitsubishi, and Lotte holding meaningful positions in specific regions and grades. That ranking matters less than what it implies: when a plant goes down, a vessel gets delayed, or a grade gets discontinued, the big producers have more levers to pull.

When you're buying 20–30 tons a year, you're a rounding error to any of them. What you're really buying is their ability to keep shipping.

I've had two supply interruptions in eight years. Both were resolved faster because our primary supplier had regional inventory and a technical rep who answered the phone. I'm not 100% sure a smaller converter would have failed us — but I know the response would have been slower.

Teijin's Panlite has a strong reputation in optical grades. If we were molding lenses, I'd look hard at it. We're not. We need consistent melt flow across large lots and a documentation package our aerospace customer already trusts. That's a different scorecard.

So when someone frames "Covestro vs Teijin" like it's a single ranking, I push back. There's no universal winner. There's a fit between your application, your qualification burden, and your risk tolerance.

The TPU vs Silicone Cutting Board Question (Yes, Really)

One of our consumer-goods customers asked us to quote a TPU overmold for a cutting board line. They were switching from silicone because of cut resistance and — this is the part buyers miss — because silicone's compression set was causing returns.

On the shelf, the silicone board looked cheaper. In our cost model, it wasn't.

Silicone tooling ran slower, the trim scrap was higher, and the return rate on cut-through complaints was around 3.5%. Not catastrophic. But at their volume, 3.5% returns plus reverse logistics ate the entire retail price gap. TPU let them run a thinner wall, faster cycle, and fewer complaints. The per-unit material cost was higher. The per-unit delivered cost was lower.

That's the whole argument in one product category.

"But Aren't You Just Paying for the Brand?"

Sometimes. I've rejected two "premium" polymer quotes in the last three years because the TCO math didn't support them. One supplier wanted a 22% premium for a grade whose only real advantage was a nicer datasheet. We didn't need the advantage. We needed parts that passed inspection.

So no, I'm not saying expensive is always better. I'm saying cheap is a claim that needs proof.

The TCO spreadsheet I use has five columns beyond unit price: scrap rate, rework hours, requalification cost, inventory carrying cost, and supply-interruption risk. When I fill those in, the ranking changes more often than not. Roughly 60% of the time — probably more, depending on how you count rework — the lowest quote is not the lowest total. That number is from my own records, not an industry study. Take it as one buyer's sample, not gospel.

What I Actually Do Now

Three rules, applied to every polymer purchase over $10,000:

None of this is glamorous. It's spreadsheet work and awkward conversations with suppliers who'd rather talk about price per kilo.

The point isn't to avoid cheap materials. The point is to stop letting the invoice's first number make the decision for you.

Buy the outcome. Negotiate the resin.