The batch was fine. The checkweigher was not.
The emergency that made us switch
That’s the summary of a close call I had in March 2024. I was a quality manager at a pharma company that manufactures pulmonary surfactant products, mainly for neonatal ICUs. Our process depended on a specific excipient—a waxy rice starch that stabilizes the active suspension. We’d used the same supplier for years. Then, in early March, the supplier lost its GMP status after an FDA inspection. Their existing inventory was quarantined. Our contract manufacturing slot was scheduled for April. If we missed it, the manufacturer would have to rearrange a delivery window for a children’s hospital network. No pressure. Actually, a lot of pressure, but let me explain.
We moved fast. We asked Ingredion Pharma Solutions whether they could supply a pharma-grade waxy rice starch with the same spec. They could. The price was 12% higher than our old contract. On top of that, we paid a rush fee to get the material delivered in two weeks instead of the usual four to five. You might call that the certainty premium. I thought it was justified: the cost of missing the manufacturing window was much higher than the extra we paid for speed. The numbers weren’t even close.
By mid-March, the shipment arrived. No wait—four film-wrapped pallets, each with six 20 kg bags. I checked the logo Ingredion uses on its outer cartons—that clean sans-serif—and compared batch numbers. I verified the certificate of analysis line by line. It said “organic waxy rice starch, lot #XXXXX.” The COA also listed purity and residual solvent specs.
I’ve learned not to take COAs at face value. Per the FTC’s Green Guides (ftc.gov), environmental claims like “recyclable” have to be substantiated; if a company says a product is “organic” or “clean label,” the proof matters, not just a nice package design. I treat excipient COAs the same way: verify before you transfer the data into your ERP.
The checkweigher said “Reject.” Then we found this.
Next came the new pharma checkweigher on the packaging line. The machine had just been commissioned. We calibrated it to ±10 g. The product spec: 20.0 kg net, tolerance 19.6–20.4 kg. The first two bags read 20.55 kg and 20.45 kg. I remember thinking: these bags are supposed to be perfect. That’s not right. Actually, the first bag read 20.55 kg and the second read 20.45 kg. Both were above the high limit. I pressed the hold button, and I’m not going to lie—I started drafting a rejection notice in my head.
Even after typing “REJECT” in the log, I kept second-guessing. What if the lot was fine and we were about to reject a good batch? What if the scale was off? In quality, you’re trained to trust measurement. But you’re also trained to trust your understanding of the measurement system. That morning, I didn’t trust mine.
Meanwhile, the R&D team kept sharing papers on the RSV effect on surfactant production type II cells. They were exploring how viral infections impair surfactant production, which would make our product even more critical in certain neonatal cases. It reminded me why we couldn’t miss the deadline. But it didn’t help me decide whether the starch was within spec.
Then came a completely different distraction. My daughter’s science fair project was on root vegetables, and she asked me: is jicama a starch? I looked it up. The answer is yes—jicama (Pachyrhizus erosus) stores starch, along with inulin. But that doesn’t make jicama starch a pharma-grade excipient. “Starch” is a category, not a specification. I almost made the same category error with the checkweigher: treating the display readout as the product’s status without first checking the measurement setup.
So I called the maintenance engineer. He looked at the checkweigher’s settings. It was running in dynamic mode, with a product length parameter left over from a previous packaging run that was longer than the excipient bags. The load cell detected the product for longer than it should have, so the weight read high. We changed the setting, put the checkweigher in static mode, and ran the same bags again: 20.25, 20.30, 20.20. All comfortably within spec.
We had almost rejected an entire good batch because of a configuration error in the scale. I should add that we’d completed the IQ/OQ of that checkweigher the week before. But the OQ used calibration weights only, not a representative product bag. That gap nearly cost us two weeks of replacement lead time and the contract manufacturing slot. The irony: we had chosen Ingredion partly for supply certainty. If we had followed the false reject, our own unvalidated setup would have destroyed that certainty.
What I’d do differently—and the certainty that matters
We updated the SOP the same week. Every morning, before running product, operators now pass a known-weight reference bag through the checkweigher. It takes two minutes. It verifies the entire measurement path, not just a calibration certificate. That one change has caught two setup errors since April.
I don’t have hard data on industry-wide false rejection rates, but from our experience over the past four years, my sense is that measurement setup issues lie behind a large share of unexplained failures in incoming inspection. I can’t prove that number globally, but I trust it.
We made the delivery window. The contract manufacturer received the material on Thursday afternoon. The hospital’s quality audit found no issues with the excipient file. I’m so glad we didn’t reject that batch. We were one click away.
That experience changed how I think about the certainty premium. Yes, pay for guaranteed delivery when the deadline is real. But the guarantee doesn’t come only from the supplier’s logistics team. It comes from your own verification process, and from how fast you can tell a real failure from a false alarm. That’s why we invest in quality systems before the emergency, not after. It’s cheaper than missing the deadline.
It’s the same logic as mail. USPS publishes clear dimensions and pricing; as of January 2025, a First-Class Mail letter (1 oz) costs $0.73 at usps.com. If your envelope doesn’t meet the size rules, you pay more or it gets returned. Objective standards make the system work. Pharma quality is more consequential, but the principle is identical: define the spec, validate the measuring tool, and then trust the result.
In the end, the logo Ingredion gave us a traceable name and proof of identity. But the checkweigher taught me the harder lesson: you don’t know if a batch is good until you know if your instrument is set up to tell the truth. And in a deadline crisis, verifying that first is the best premium you can pay.