Steinert NIR vs. Conventional Sorting: A Purchasing Administrator’s Comparison

What I’m comparing — and why

I’m the purchasing administrator for a 180-person recycling and aggregates company. I manage about $1.4 million in annual spend across 14 vendors, and I report to both operations and finance. When I took over purchasing in 2021, I thought sorting equipment was a technical decision. By 2024, I’d learned it’s also a brand decision — because what leaves our facility is what our customers judge us on.

This comparison is narrow: Steinert NIR sensor sorting vs. conventional manual and density-based sorting for mixed recyclables. I’m not comparing Steinert to a specific competitor, and I’m not saying one option wins everywhere. The three dimensions I use are output quality, uptime and maintenance, and total procurement cost. Those are the ones that get me called into meetings.

Dimension 1: Output quality and customer perception

With conventional manual sorting, the upside is flexibility. Pickers can catch odd items that a sensor might reject. The downside is consistency. On a Monday morning crew vs. a Friday afternoon crew, the same infeed can produce different bale quality. I didn’t fully understand that until a March 2023 customer audit. A buyer rejected two loads because PET bales had too much PVC. We lost the order and, worse, we looked sloppy.

Steinert NIR sensor sorting uses near-infrared spectroscopy to identify materials by their spectral signatures, according to Steinert’s published equipment information (steinertglobal.com). In practice, that means it can separate polymer types — PET, HDPE, PP — more consistently than manual sorting, assuming the feed is prepared and the calibration is right. It isn’t magic. Black plastics and heavily contaminated streams can still be difficult.

According to the European Commission’s Circular Economy Action Plan (ec.europa.eu, 2020), better sorting is central to higher recycling rates. That’s policy language, but it matches what buyers ask for: cleaner streams and fewer surprises in the bale.

The comparison conclusion: if your customers audit bale quality or you sell into higher-grade end markets, NIR usually gives you a more repeatable output. If your stream is low-volume and highly variable, manual sorting can still be the more practical choice. Quality is a brand signal. When I switched one line from manual QC to a stricter spec, customer complaints dropped, but I had to explain the added cost to finance. The point wasn’t premium for its own sake; it was avoiding the cost of looking unreliable.

Dimension 2: Uptime, maintenance, and operational reality

I assumed “NIR-ready” meant the same thing across vendors. Didn’t verify. Turned out some systems need different air, power, and cleaning routines. That mistake cost us two weeks of commissioning time (unfortunately).

Conventional sorting has fewer electronics, so when it breaks, it’s often obvious and easier for our team to fix. NIR systems need clean optics, stable compressed air, and someone who actually understands calibration. If you don’t have that person, uptime suffers. The surprise wasn’t the technology cost. It was how much the service model mattered. A fast-response technician and a clear spare-parts list were worth more than a slightly lower purchase price.

On the floor, the frustrating part is the same issue recurring: operators bypass a sensor because it “acts up,” then quality drops. You’d think a written SOP would prevent that, but shift pressure is real. What finally helped was a simple dashboard that showed reject rate by shift. That made the problem visible without blaming anyone.

Comparison conclusion: NIR wins on consistency when maintained; conventional wins on repairability in low-tech environments. The deciding question isn’t “which is better?” It’s “who on my team will own calibration and cleaning every single day?”

Dimension 3: Procurement cost, TCO, and the finance conversation

I’m not going to quote equipment pricing here. If someone gives you a firm number without a site survey and feed analysis, be skeptical. What I can compare is cost structure.

Conventional sorting is lower capital cost, higher labor exposure. You pay every shift, and turnover hits quality. NIR is higher upfront, but it can reduce labor on specific streams and create more consistent bales. The hidden cost is maintenance and training. In our 2024 vendor consolidation project, we bundled service agreements and spare parts into the evaluation. That changed the ranking. The cheapest quote wasn’t the lowest total cost once we added calibration visits and downtime risk.

For compliance and finance, the invoice matters too. A vendor who can’t provide clear equipment specs, warranty terms, and service invoices creates problems. I once ate $2,400 in rejected expenses because a supplier provided handwritten receipts. Now I verify invoicing capability before any order. That’s boring, but it’s real procurement.

Comparison conclusion: choose NIR when throughput and quality justify the operating discipline. Choose conventional when volume is low, labor is stable, and capital is tight. Don’t assume the expensive option is automatically better — or that the cheap one is low quality. Match the tool to the stream.

Which option fits your situation?

Here’s how I’d decide now:

• High-volume mixed plastics, customer audits, and end markets that penalize contamination: test Steinert NIR sensor sorting with a pilot.
• Low-volume, highly variable waste, limited technical staff: conventional sorting may be more realistic, with better manual QC.
• Ferrous and non-ferrous metals: magnetic separation and eddy current systems are often the right first step; NIR is not a replacement for that.
• Budget-constrained but quality-focused: phase the project. Start with the stream that causes the most customer complaints, then expand.

If you searched for Hilla Steinert, Millennium Lego, Eddie Outlet, or “what is skiing versus downhill skiing?”, those terms aren’t related to this article. I can only speak to the industrial sorting side.

What I’d do differently next time

I’d demand a site-specific trial, not a lab demo. I’d ask for the maintenance schedule in writing — or rather, I’d ask for the real schedule, not the sales version. I’d talk to an operator who runs the same system on a similar feed. And I’d check whether the vendor’s service team can actually get to our site within a day.

Quality isn’t just about the machine. It’s about what your customers see when they open the bale. That’s the comparison that matters.