FNSKU Mis-Scans at Amazon FCs: The Print Quality Problem Most Prep Centers Won't Discuss

By Bernardo Campelo — Forbes Business Council E-Commerce Leader, Amazon SPN Certified provider, Amazon SP-API authorized partner, and Founder of PrepVia.

I get the same email every quarter. A seller pulls up his Seller Central, screenshots the inbound fees, and asks me the same question: "How is this possible? The FNSKU labels were printed. The prep center confirmed every unit was labeled. Amazon received the shipment. So why am I being charged $1,200 in defect fees for unscannable barcodes?"

It is a fair question. And the answer almost never comes from the prep center.

The honest answer is that the label was probably there. It was probably the right FNSKU. It just was not scannable when the Amazon fulfillment center scanner tried to read it. That happens for a specific, mechanical reason that most prep centers either do not understand or do not want to discuss with their clients.

I run PrepVia, an Amazon SPN Certified prep center that has shipped tens of millions of FNSKU-labeled units. We audit our own label output every quarter, and what we have found about print quality is not what the industry talks about. This is the part of the FNSKU conversation that costs sellers real money, and almost nobody is writing about it.

The Invisible Tax on Every Shipment

Amazon's inbound defect fees are not new. What is new is how aggressively they are being applied. In 2026, fees for unscannable barcodes run $0.32 to $5.72 per unit depending on category and size. That fee is charged regardless of whether the seller or the prep center was at fault. Amazon does not arbitrate. The barcode failed to scan, the seller pays.

This creates a strange dynamic. The prep center has no financial incentive to fix label quality because they do not pay for the failures. The seller pays. The prep center never sees the fee on their invoice. They see a screenshot from a frustrated client three weeks later and tell them, in slightly different words, that Amazon's scanners are the problem.

Sometimes Amazon's scanners are the problem. They run faster than handheld scanners, they read at angles, they read through plastic and tape, and they have to handle 50 units per minute on a conveyor. But the FNSKU specification is published. The Code 128 standard is well known. Industrial label printers exist that meet the spec at any volume a prep center will ever run. The "it is Amazon's scanner" answer is sometimes true and almost always incomplete.

What is usually true is that the prep center is printing labels that work in their warehouse and fail at the FC.

Why a Valid FNSKU Fails at the Fulfillment Center

The FNSKU number is correct. The barcode encodes the right data. The label has the human-readable text below it. By every check the prep center runs, the label is "good." Then it gets into a sealed box, travels 1,200 miles in a freight truck, sits on a conveyor at an FC for three seconds, and the scanner cannot read it.

Here is what is actually happening at that scanner moment.

Resolution that looks fine but is not

This is the most common failure mode. The FNSKU rendering library generates a PNG. If the PNG resolution is too low, the printed barcode has fuzzy edges that look acceptable on a phone screen but produce ambiguous transitions for a high-speed industrial reader. A barcode reader works by measuring the width of each bar and space. Fuzzy edges introduce measurement noise. Above a certain noise threshold, the decoder gives up.

When we audited our own label output at PrepVia last quarter, we found this exact problem. Our rendering library was generating barcode PNGs at a scale that looked sharp on a 1080p monitor but came out marginally fuzzy on the thermal printer. We doubled the scale parameter and doubled the vertical bar height. Visually, the label changed slightly. Functionally, first-scan success at downstream readers went from a number we were embarrassed by to one we will publish.

That change cost us nothing in materials or speed. It is a number in a configuration file. But it was off, for a long time, because nobody had audited it. The labels "scanned fine" inside our facility.

Thermal printers that drift out of calibration

Industrial thermal printers heat tiny elements on the print head to burn the barcode pattern onto a label. Over time, the heating elements wear unevenly. The print darkness drifts. A printer producing crisp barcodes in March might be producing washed-out barcodes in November.

Manual prep centers rarely calibrate. They print, they ship, they wait for complaints. The complaints arrive as "Amazon scanner problems" and the printer keeps drifting.

The fix is mechanical. Every thermal printer should be calibrated against a known-good reference at a set interval, and print density should be measured, not eyeballed. This is operational discipline most prep centers do not maintain because it costs labor that does not generate billable revenue.

Inkjet and laser printers being used for industrial labeling

Some prep centers still use inkjet or laser printers for label printing. Inkjet ink fades. Laser toner smudges under heat and pressure. Both technologies were not designed for warehouse environments where labels sit in sealed boxes for weeks, ship in trucks that hit 110°F, and arrive at FCs that handle them roughly on conveyors.

Thermal is the industry standard for industrial labeling for a reason. If your prep center is using anything else, the labels are at higher risk regardless of how good they look at the moment of printing.

Labels that do not fully cover the old barcode

The product almost always arrives at the prep center with an existing barcode on the packaging — a UPC from the manufacturer, a previous seller's FNSKU, or both. If the FNSKU label is applied without fully covering the previous barcode, the Amazon scanner can read either one. When it reads the wrong one, the unit is checked into the wrong account or rejected as a mismatch.

The fix is operational, not technical. Cover 100% of the old barcode, always, no exceptions. Some prep centers leave a corner exposed to "save sticker stock" or make label removal easier for the customer. Both are false economies measured in inbound defect fees.

Adhesive failure in heat and humidity

A label that does not stay on the product is not a barcode problem in the strict sense, but the result is the same. No barcode to scan. Cheap prep-center labels use general-purpose adhesives that fail in heat or humidity. A pallet sitting in a non-air-conditioned dock yard in Texas in July loses labels at a measurable rate. Permanent adhesives rated for industrial logistics environments cost slightly more per label and eliminate the failure.

The Math of a 3% Failure Rate

Take a shipment of 5,000 units to Amazon. The prep center reports 99.95% scan accuracy in their internal system. The label was printed, applied, and scanned before pack. By every internal check, the work is done.

The shipment arrives at the FC. Amazon's scanners run at conveyor speed. Of the 5,000 units, 4,850 scan first try, 130 require a second pass, and 20 are flagged as unscannable and routed to Amazon's manual prep team. The unplanned services fee on those 20 units runs $1.50 to $2.50 each in most categories. The seller is invoiced $30 to $50 for that shipment.

Multiply across 12 shipments per month. Multiply across a year. The seller is paying $4,000 to $8,000 annually for label failures the prep center never reported because their internal scanner accepted the labels.

How to Audit Your Prep Center's Label Quality

Sellers can test label quality directly. The audit is simple, and the prep center does not need to know it is happening.

2D Barcodes: The Industry Is Moving On

The FNSKU is still encoded in Code 128, a 1D barcode standard that has been around since 1981. It does the job. But 1D barcodes have known limits: they need a clear line for the scanner to cross, they are vulnerable to localized damage, and they cannot encode much data in a compact space.

2D barcodes — DataMatrix is the most common in industrial applications — solve all three. They scan from any angle. They tolerate partial damage through built-in error correction. They encode roughly four times more data in roughly one-quarter the physical space. They are standard in pharmaceutical labeling, aerospace parts tracking, and most modern manufacturing.

At PrepVia, we now print DataMatrix codes on every internal tracking label inside the warehouse, for inventory location, audit trail, and operator scanning workflows. Scan speed on mobile devices is roughly twice as fast as Code 128. The built-in error tolerance means we can keep using a label even after it has been bumped or partially obscured by tape.

Amazon's FNSKU is still Code 128 today, so customer-facing FNSKU labels stay 1D. But the direction the rest of industrial logistics is moving is unambiguous. The prep centers that are already operating with 2D-capable workflows will have a head start when Amazon eventually updates the spec. The ones still hand-printing labels on inkjet will not.

What the FC-Quality Operation Looks Like

The fix for all of this is not a single thing. It is operational discipline across five mechanical points.

None of this is glamorous. None of it shows up in a marketing brochure. It is the boring operational work that separates prep centers that work from prep centers that look like they work.

Why This Problem Persists Across the Industry

When I think about why this issue keeps showing up across thousands of prep centers in the United States, the answer is structural. Prep centers do not pay for the failures their labels cause. The seller pays. The fee shows up on the seller's Amazon invoice, weeks later, in a line item that does not mention the prep center by name. There is no feedback loop.

Most prep centers will never improve label quality because there is no incentive to improve it. The labels look fine in their warehouse. The shipments leave on time. Clients sometimes complain about Amazon fees, but those complaints get blamed on Amazon. The prep center keeps charging, keeps shipping, and keeps producing labels that work in the prep facility and fail at the FC.

The fix is not expensive. It is not exotic. It is a configuration change, a calibration schedule, and a willingness to measure the metric that matters instead of the metric that is easy. But until prep centers either feel financial pressure from sellers or hold themselves accountable to their own standards, most will keep doing what they have always done.

The sellers who notice this and switch are the ones who stop seeing inbound defect fees on their invoices. The ones who do not notice keep paying.

Frequently Asked Questions

Why does Amazon charge a fee for unscannable barcodes if the label is technically correct?

Amazon's receiving systems are optimized for throughput. When a label does not scan first try, the unit is routed off the main conveyor and handled by an exception process — sometimes by Amazon's unplanned services team, sometimes by manual reprinting. That process costs Amazon labor. The fee recovers that labor, regardless of whether the label was "technically correct" by the original print spec.

Can I appeal an inbound defect fee for unscannable barcodes?

You can open a case with Seller Central, but appeals on barcode quality are almost never granted. Amazon's position is that the seller is responsible for label quality regardless of who actually printed the labels. The only effective remediation is changing prep centers.

What is the actual industry standard for FNSKU label quality?

Amazon's published spec is Code 128 at sufficient resolution to scan at conveyor speed. There is no single minimum DPI in the spec, but industrial practice is 203 DPI minimum on a properly calibrated thermal printer, with a quiet zone of at least 1/8 inch around the barcode and a print contrast ratio above 90%. Most prep centers exceed the spec on a freshly serviced printer. Many drift below it within weeks.

How often should a prep center recalibrate its thermal printers?

Industry best practice is monthly verification against a reference standard. At PrepVia we calibrate weekly because our volume puts more wear on the heads. Prep centers without a calibration schedule are operating on assumption.

Is DataMatrix really better than Code 128 for FBA labels?

For FBA customer-facing FNSKU labels, Amazon's current spec is Code 128, so you cannot use DataMatrix there today. But for any internal tracking — receiving, location, audit trail, operator workflows — DataMatrix is faster to scan, more damage-tolerant, and more space-efficient. Most modern industrial logistics has moved to 2D.

How can I tell if my prep center is using inkjet or laser instead of thermal?

Look at the label closely. Thermal labels have no ink — the print is burned directly into heat-sensitive label stock. Scratch the label gently with a fingernail. If anything smudges or transfers, it is inkjet or laser. If nothing comes off, it is thermal. Thermal labels also age differently — they may darken or fade gradually but they do not smear.

Does the FNSKU label position on the product matter for scanning?

Yes, significantly. Amazon's spec requires the FNSKU to be visible and scannable from at least one major face of the product. Best practice is to apply the label on the same face as the original UPC, covering the UPC entirely, and avoiding seams, curves, or transparent surfaces that distort the barcode shape.

Final Take

The FNSKU label problem looks small. It is a $0.20 sticker. The cost of fixing it is a configuration change and a calibration schedule.

Scale it across 50,000 units per month, twelve months a year, and the same problem becomes thousands of dollars in inbound defect fees that almost no seller traces back to label print quality. The fee is invisible. The cause is invisible. The prep center has no incentive to investigate.

Most things in operations are like this. The big losses come from small mechanical details nobody owns because no single person sees the full cost. The work of running an actual logistics operation is finding those details and making someone responsible for them.

The labels at the bottom of a box arriving at an Amazon FC are one of those details. If your prep center has never talked to you about FC scan rates, print resolution, or thermal printer calibration, they have not done the work. Whether that costs you $500 a year or $50,000 depends on your volume. But it is costing you something.