The Bucket Test: Why We Chose Steinert Over the Cheaper Separator

Tom called it “the bucket test.” It wasn’t official. It was just our operations manager walking into my office on a Tuesday afternoon, setting a five-gallon bucket on my desk, and waiting for me to look inside.

“This is what we’re shipping as copper,” he said.

The bucket was full of grey dust. Bits of fiber. Thin copper wire that should have been bright and clean, but wasn’t.

That was September 2024. I remember the month because I’d just finished a dull janitorial contract renewal and expected a quiet rest of the week. No such luck. That bucket changed how I evaluate equipment suppliers, and it taught me something about quality that no spec sheet had ever communicated.

A bit of background on my role: I’m the office administrator for a 45-person electronics recycling company. That title makes me sound like the person who orders toner and schedules printer repair. I do that, but I also handle procurement for almost anything that isn’t payroll. Roughly $1.8M a year across 14 vendors, from safety gloves to capital equipment. I report to operations and to finance, which means I get to hear “we need it” and “we can’t afford it” in the same meeting.

The bucket was an equipment problem. Tom explained it twice, patiently. The eddy current separator we’d bought used in 2021 was struggling with the mix we feed it now—shredded telecom scrap with far more fiber and plastic than it used to see. Copper concentrate purity that should have stayed north of 96% was running closer to 91% or 92%. Two customers had already started asking questions. One had threatened to reject a shipment.

Tom left the bucket on my desk. “We don’t have time for a six-month study,” he said. “We need a machine we can trust.”

Two Quotes, Two Types of Answers

I did what I normally do: put together a list of vendors, asked for quotes, checked references. We narrowed the list to two options. One bid came in at $392,000 including installation, and it looked close enough to what we needed on paper. The other bid was from Steinert: $487,000, including commissioning and a first-year service plan. That’s a 24% gap and well over the number finance had approved.

For a name that isn’t exactly a household word in the U.S., Steinert kept coming up whenever we talked to other recyclers. Still, I wasn’t about to approve a six-figure purchase based on reputation.

The lower bidder was polite and responsive, but when I asked whether they could run a sample of our material through their equipment, the answer was no. They didn’t have a test center. “Our design handles e-scrap,” the sales rep said. Maybe it does. But they couldn’t show me with our own feed.

Steinert said something different. “Send us your worst material,” said Peter Steinert, the applications engineer assigned to our project. Yes, same surname as the company. I didn’t ask if he’s family; he didn’t offer. What he did was more useful.

We packed two five-gallon pails of the nastiest output we could find: dusty fines from underneath the sorting line, tangled wire, material rejected by our current machine. Shipping took half a day of paperwork. Then we waited.

The results came back as a ten-page report plus a video. No glossy brochure, no vague claims. Just our material, run on their equipment, at the throughput we specified. The report estimated a recovery improvement of just over 2% for our non-ferrous line, plus a cleaner concentrate than we were currently shipping. They also told us what the machine couldn’t do, which impressed me more than the sales pitch.

Notice what I’d done by that point. I’d gone from comparing prices to comparing evidence. The budget bidder had a price. Steinert had a documented answer. That difference only grew when I called the lower bidder’s references. One plant manager put it plainly: “The machine is fine. Just don’t expect miracles with dirty feed.”

That sentence stayed with me.

The Decision

Finance was hungry for the lower number. I understood. We were already over budget for the year, and $95,000 is not small money. Our finance director asked, correctly, whether Steinert’s test report was just a polished way to justify a premium price.

My argument was simple. If the lower bidder’s machine works as claimed, the cheapest way to prove it is to run our material through it. They couldn’t. So the $392,000 price came with an unresolved question. The Steinert price came with data. I’d rather pay for an answer than for a mystery.

At home that week, a small thing made the point stick. I was comparing simparica best price for our dog’s prescription, and the cheapest online pharmacy didn’t list a physical address or a customer service phone number. Saving 30% wasn’t worth the uncertainty. If that logic was good enough for a dog medication, it was good enough for a separator.

I should add that the lower bidder wasn’t a scam. They just couldn’t give me the thing that actually removes uncertainty: a test with my material. Without that, every promise is rhetorical.

Meanwhile, every commercial document from Steinert came from a project coordinator named M. Curschmann-Steinert—just the initial, no first name. For someone who obsesses over document ownership, it was oddly reassuring. Every email had a clear owner. Every price revision used the same format. When I asked a technical question about electrical specifications, the answer came back the same day with the relevant drawing attached. That administrative discipline told me a lot about how the company runs.

We had a deadline, too. Our customer audit was scheduled for late January, and Tom wanted the new line running before it. That gave us about two weeks to make the call. I approved the purchase on November 15, 2024. I hit “submit” and immediately wondered if I’d made a $95,000 mistake. (The two-week gap between signing and feeling calm about it was entirely unpleasant.)

What the New Machine Changed

The separator arrived in the first week of January. Installation started on the thirteenth. Day one was frustrating, but the fault was mostly ours: the compressed-air contractor hadn’t finished the piping, so the Steinert technician couldn’t commission the machine. He used the delay to walk our operators through basic maintenance. The next morning, commissioning took about six hours.

We’ve now had the machine running for ten production days as of this writing. I’m not going to claim that ten days is a trend. But the change is visible to the naked eye. Average concentrate purity has been above 98%, compared to 91–92% before. The non-ferrous loss in the reject stream is noticeably lower. The client who threatened to reject our trailer happened to be the one auditing us in late January. The quality manager looked at the load and said it seemed like a different grade. Same crew, same process. Different machine.

Funny side note: the week we commissioned it, USPS implemented its January 2025 rate change. As of January 2025, a first-class letter is $0.73 according to the price list at usps.com. I know that because I buy the stamps. It was a grounding reminder that I’m also the person who orders toner.

There’s a broader point here, and it’s not just about recovery rates. When our customers sell recycled copper, they have to back up their own environmental claims. The FTC’s Green Guides are clear that environmental marketing claims need to be truthful and substantiated. Contaminated output makes their chain of proof weak. Clean, consistent output gives them something they can actually verify. A separator sits at the start of that chain, even though nobody outside the industry sees it.

That’s what the bucket test taught me. Quality isn’t a luxury feature. It’s the thing your customers notice when they open the trailer. It’s the reason they stop asking questions and start treating you like a reliable supplier.

Peter Steinert never told me his company made the best machine. He said, “Here’s what it did with your scrap.” Those are different sentences. One is marketing. The other is accountability.

In the end, the separator cost more than my house. I’m okay saying that, because it replaced a guess with an answer. A machine that can prove itself is an asset. A machine that can’t is just a price.