📅 Prepared May 18, 2026
✍️ Dirk Adams
⌛ 12 min read
FAT Research Series — USDA Labeling Rules / Retail Labels & Grocers
USDA Food Safety and Inspection Service regulations require every federally inspected meat, poultry, and catfish package to bear the inspection legend and the establishment number, prominently and legibly. In practice, the establishment number routinely lands on package seams, glued flaps, and low-contrast back panels. The placement is driven by variable-data printing economics — not by any technical barrier to compliance.
This paper was prepared with the assistance of AI.
FAT thesis: The regulatory text already requires legible, prominent placement of the establishment number. The “elsewhere on the labeling” provision at 9 CFR 317.2(c)(2)(ii) is a permissive option for the processor, not a license to bury required disclosure. The per-package cost of bringing the label into compliance is immaterial — single-digit cents fully amortized, falling to a fraction of a cent over a normal print-head refresh cycle.
Thesis
USDA Food Safety and Inspection Service (FSIS) regulations require every federally inspected meat, poultry, and Siluriformes (catfish) package to bear two pieces of mandatory disclosure: the official inspection legend (the round shield) and the establishment number. The regulation is unambiguous that this information must be “prominent, conspicuous, and legible.” In practice, a substantial portion of supermarket meat and poultry labels print the establishment number on package seams, the back of the package, under glued flaps, at low contrast against the product, or at type sizes near the regulatory minimum — placements that satisfy the letter of a permissive sub-clause (9 CFR 317.2(c)(2) “elsewhere on the labeling”) while undercutting the consumer-facing intent of the surrounding rule (9 CFR 317.2(b) “prominent and conspicuous”).
The placement choice is not arbitrary. It is driven by the mechanical reality of how meat is packaged at industrial scale: pre-printed roll-stock film purchased weeks in advance from third-party converters, with variable data (lot code, sell-by date, weight, establishment number) added at the end of the production line by inkjet print heads that are physically located at the seam. Compliance — moving the establishment number to a position adjacent to the inspection legend on the front panel — is achievable but requires capital investment, plate revisions, and SKU rationalization that the regulatory regime does not currently force.
This paper documents the regulation, the mechanical explanation, and a defensible estimate of the per-package cost of full compliance. It is the first in a planned FAT Research series on the operational drivers of meat label disclosure failures.
1. What the regulation actually says
1.1 The two mandatory elements
9 CFR 317.2(c)(1) (meat) and 9 CFR 381.96(a) (poultry) require every label bearing the official inspection legend to carry the establishment number assigned by FSIS to the plant where the product was prepared.
1.2 Where the number can go
9 CFR 317.2(c)(2) — emphasis added:
“The official establishment number shall be shown … (i) Within the official inspection legend in the form ‘EST. (number)’; or (ii) Elsewhere on the labeling, provided that the manner of showing the number is approved by the Administrator, and that the number is prominently and legibly applied.”
The corresponding poultry rule at 9 CFR 381.96(b) reads materially the same. So three placements are technically permitted, in declining order of legibility:
- Inside the round shield (the 1906-era convention)
- Adjacent to the legend
- Anywhere else on the labeling, as long as the placement is FSIS-approved and the result is “prominently and legibly applied”
Option (3) is the loophole this paper examines.
1.3 The conspicuousness and legibility rules
The placement flexibility in 317.2(c)(2)(ii) does not unwind the surrounding rules. Two parts of the same section govern how the number, wherever placed, must actually appear:
9 CFR 317.2(b):
“All information required to be shown on the labels … shall be in lettering of such size, color, and style and so placed thereon as to be likely to be read and understood by the ordinary individual under customary conditions of purchase and use.”
9 CFR 317.2(h):
“Information required to appear on a label … shall appear … in print or type of a uniform size and conspicuousness … not less than [the minimum specified size].”
The “ordinary individual under customary conditions of purchase” is the regulatory yardstick. A consumer reaching for a tray of ground beef in a refrigerated case has not been granted X-ray vision through a glued flap.
1.4 What enforcement looks like in theory
Labels are approved by FSIS’s Label Submission and Disclosure Specialist (LSDS) program before they appear on product. After approval, in-plant FSIS inspectors verify that the label is being applied — not that placement still reads well at the supermarket case after weeks in distribution and a store-applied price sticker.
Non-compliance Records (NRs) and Memoranda of Interview (MOIs) issued at the plant are the enforcement instruments available after approval. NRs cite the specific regulation violated (the regs field in FSIS Inspection Task data). The question for the section below is how often that regs field reads “317.2” or “381.96” — that is the empirical question of how often labeling compliance is enforced in the field, separate from label-approval review.
2. How often is labeling actually cited?
Methodology. A random sample of 500 FSIS establishments from the publicly distributed FAT FSIS extract was scanned. For each establishment, the publicly available administrative-action records (NRs and MOIs from FSIS Inspection Task data, Steps 4 and 5 of the FAT monthly FSIS pipeline) were inspected for the regulation citations 317.2, 381.96, and 412.x, plus keyword matches on “establishment number,” “legibility,” “inspection legend,” and “official mark.”
Result of the public-extract sample. The publicly available data is far thinner than what FSIS holds internally:
| Metric | Count in 500-establishment sample |
|---|---|
| Total NRs surfaced in public extract | 4 |
| Total MOIs surfaced in public extract | 3 |
| NRs citing 317.2 / 381.96 / 412.x | 0 |
| MOIs citing 317.2 / 381.96 / 412.x | 0 |
The top regulations cited in the small surfaced NR pool were all 9 CFR 313.x (humane handling at slaughter) — 313.1, 313.2, 313.15(a)(1), 313.16(a)(1). Zero label-regulation citations were observed.
What this finding does and does not support. The point of the count is not to assert a population number. Four observed NRs across 500 establishments is too thin to draw a conclusion about the underlying enforcement rate — the public FSIS extract clearly under-reports relative to what plant inspectors actually issue. The honest finding is twofold:
The publicly available NR/MOI extract is not dense enough to answer the question. The 51,858 GCP and 102,517 LHH task rows reported in the monthly pipeline are largely successful task observations, not NRs. Only a small subset becomes an NR (469 NRs total in the last LHH run across the entire FSIS inspected universe). At that rate, 500 random establishments will surface a handful of NRs at most.
In the data we can see, label citations are absent. Among the NR/MOI text public consumers can read, FSIS labeling rules are simply not visible. Whether that reflects under-citation by inspectors, under-publication by FSIS, or genuine compliance is not determinable from the public extract alone.
What would actually answer it. A targeted FOIA request to FSIS for the disposition of NRs citing 317.2 / 381.96 over a defined period (e.g. FY 2023–2025), broken down by violation type (“placement,” “legibility,” “missing establishment number,” “missing inspection legend”) would produce a defensible numerator. This is the next step in the research program.
The “rarely enforced” intuition stated in earlier FAT editorial commentary is therefore neither confirmed nor refuted by the public data available today. FAT should retract that specific framing pending FOIA results and instead frame the case from the regulatory text and the mechanical analysis that follow.
3. The mechanical explanation — variable-data printing on flexible packaging
The remainder of this paper focuses on why the establishment number ends up on seams and back panels — independent of how frequently FSIS chooses to cite the placement. The mechanical answer is well-established in the flexible-packaging industry.
Definition — what is a “converter”?
A flexible-packaging converter is a third-party industrial printer that does not make the underlying plastic film and does not pack meat. The converter sits in the middle of the supply chain. A film extruder produces the raw multi-layer plastic in long blank rolls; the converter receives those rolls, prints multi-color graphics onto them on a high-speed flexographic or rotogravure press, slits and re-wraps them to the meat packer’s specifications, and ships finished branded roll-stock to the packing plant. The packer loads the finished roll onto a form-fill-seal or thermoform line, which wraps the film around the product on the plant floor.
The verb “convert” in this trade means “convert blank film into branded packaging.” The converter is the printer; the meat packer is the customer; the meat packer’s own line does only minimal printing (the variable data described below). When the paper refers to “the converter,” it means this specialty printing company, not the brand and not the packer.
Major U.S. flexible-packaging converters serving the meat industry include Sealed Air’s Cryovac line, Berry Global, ProAmpac, Winpak, Amcor (formerly Bemis), Sonoco, and Constantia Flexibles.
3.1 The two-stage print model
Modern meat and poultry packaging is built in two distinct print stages.
Stage 1 — pre-printed roll stock at the converter. The bulk of the label graphics — brand mark, food photography, welfare and quality claims, nutritional panel, recipe suggestions, allergen statement, UPC, the round USDA shield design itself — is printed onto a continuous roll of flexible film by the converter, weeks before the film reaches the plant.
Converter print runs are large: typical economic minimums are in the range of 50,000 to 500,000 linear feet per design, run on flexographic or rotogravure presses with engraved plate cylinders. Setup cost for a single SKU at this stage is dominated by plate or cylinder engraving — defensible industry figures place this at $3,000–$5,000 per color for flexo plates and $1,000–$3,000 per color for gravure cylinders [VERIFY — industry vendor public quote-sheets vary; cite Flexographic Technical Association rate cards if cited as authoritative]. A multi-color meat label typically uses 6–10 colors. Set-up cost per SKU is therefore in the $18,000–$50,000 range, amortized across the run length. A 100,000-linear-foot run at that setup cost is roughly $0.18–$0.50 per linear foot of set-up overhead before the consumables cost of substrate and ink.
This cost structure has two implications relevant to placement:
The converter has a strong economic interest in producing plant-agnostic film. A run that bakes “EST. 12345” inside the printed shield can only be loaded at plant 12345. A run that leaves the shield interior blank (or carries only the standard “U.S. INSPECTED AND PASSED BY DEPARTMENT OF AGRICULTURE” wording) can be loaded at any plant in the brand’s network. The first model multiplies SKUs, inventory, and set-up cost; the second consolidates them.
Once the shield is pre-printed without the number, the number has to be added downstream — at the plant, on the line, after the product has been weighed and packed.
Stage 2 — variable-data overprint at the plant. The information that varies per package — lot code, sell-by date, net weight, and the establishment number itself when not pre-printed — is added at the end of the production line by an industrial variable-data print system. The dominant technologies are:
- Continuous inkjet (CIJ). Heads from Videojet, Domino, Linx, and Markem-Imaje, projecting a stream of charged ink drops onto the substrate. Print resolution is moderate (typically 80–200 dpi) and the technology handles curved, irregular, and seam surfaces well.
- Thermal-inkjet (TIJ). Higher resolution (300–600 dpi) but more sensitive to substrate and angle; common for back-panel printing on a flat substrate.
- Thermal-transfer overprinting (TTO). A ribbon-based print head that lays heated ink onto a flat film section, used on form-fill-seal (FFS) lines. Best for crisp labels.
- Print-and-apply label applicators. A separate paper sticker carrying the variable data is printed and applied to the package. Common in tray-pack ground meat where the price-per-pound sticker also carries the lot, weight, and (sometimes) the establishment number.
3.2 Why the print head sits at the seam
The variable-data print head is a piece of physical equipment that occupies a fixed position above or beside the conveyor at the end of the packaging line. The line architecture dictates where that position can be:
On a horizontal form-fill-seal (HFFS) vacuum-pack line, the package is sealed along a long-edge “fin seal” or “lap seal” at the moment the film is wrapped around the product. The most common position for the variable-data print head is immediately after the seal jaws, projecting downward onto the fin-seal area as the sealed package moves past on the conveyor. This is why the lot code, weight, and (often) the establishment number appear directly on the seam: not because anyone chose the seam as the location for consumer disclosure, but because the seam is the surface the print head can physically reach without re-engineering the line.
On a thermoform-fill-seal line (the “blister-pack” style with a rigid bottom tray and a flexible top film), the variable-data print head is typically positioned over the top-film conveyor before the lidding film is sealed down, printing onto a designated flat zone of the lidding film. Placement can be more flexible here — but only if the converter’s pre-printed top film leaves a clear print zone on a visible panel. If the top-film design doesn’t leave such a zone, the variable data falls to the lidding film’s edge or back.
On a tray-pack ground-beef line, the variable-data is usually printed on a paper price-and-weight label applied by a separate print-and-apply unit. The placement of that label on the tray is dictated by the applicator’s geometry — most commonly the long edge of the lid or the front-corner of the overwrap. Whether the establishment number rides along on that sticker depends on whether the brand has chosen to overprint the EST# on that sticker or has left it embedded in the original pre-printed film.
On a multi-plant brand using pooled labels — film printed for distribution across plants A, B, and C — the establishment number string is typically a concatenation (“EST. P-1234 + P-5678 + P-9012”). String lengths of 25–35 characters do not fit inside the standard 1.25-inch-diameter inspection shield. The variable-data print station prints the concatenated string on the seam or back panel because that is the available real estate.
3.3 Why moving the EST# back to the shield is mechanically possible but operationally costly
There is no technical barrier to printing the establishment number adjacent to the inspection shield on the front panel of the package. Three changes would be required:
Converter-level redesign. Each SKU’s pre-printed film artwork would need to be revised to leave a defined clear zone next to the shield where the variable-data inkjet head can land the number. Plate or cylinder costs apply to the redesign at flexographic / rotogravure rates discussed in 3.1.
Plant-level print-station relocation or addition. The existing variable-data print head positioned at the seam either has to be moved to the front-panel print zone (requires a new mounting position with line clearance, sensor recalibration, and re-validation) or supplemented by a second variable-data head dedicated to the front-panel placement. Industrial CIJ heads cost approximately $15,000–$35,000 capital per head [VERIFY against current vendor quotes] with installation, integration, and validation roughly doubling that figure. Depreciation over a 7-year asset life gives an annualized capital cost of $4,300–$10,000 per head per year.
SKU rationalization or pre-printed plant-specific film. If the brand wants the EST# inside the shield (rather than adjacent to it), the only path is plant-specific film at the converter — increasing converter-side setup costs as the number of SKUs multiplies by the number of plants. A brand running 4 plants × 8 cuts of beef would go from 8 SKUs to 32 SKUs, with proportional inventory carrying cost.
3.4 The cost of compliance — what is one-time, what is continuing, and what does it look like next to other line capex
The cost question is the one that gets thrown back at FAT first when this topic comes up: “moving the EST# is expensive.” It is worth separating that claim into its parts.
3.4.1 One-time costs
These are incurred once when a brand redesigns its label to place the establishment number on the front panel adjacent to the inspection shield. They are not paid again every year.
| One-time item | Order of magnitude | What it actually is |
|---|---|---|
| Converter plate / cylinder redesign | $5,000–$50,000 per SKU | New flexographic plates or rotogravure cylinders engraved to leave a clear zone next to the shield for the variable-data print station to land on. Paid once to the converter at the next print-run reorder. |
| Plant-side line modification | $0–$25,000 per line | Repositioning the existing variable-data print head, or adding a second print head dedicated to the front-panel position. Includes mounting hardware, sensor recalibration, line integration. |
| Print head itself (only if adding a second one) | $15,000–$35,000 per head | Industrial continuous-inkjet head — Videojet, Domino, Markem-Imaje, or Linx — bolted to the line. Many plants already have multiple heads; the additional head is incremental, not net-new infrastructure. |
| LSDS re-approval | Negligible | FSIS Label Submission and Disclosure Specialist review of the new label artwork. Usually under 90 days, no fee. |
Total one-time outlay for a brand redesigning a single SKU across one production line: approximately $20,000–$110,000. Lower end for a brand that already has spare variable-data print-head capacity. Higher end for a brand that has to add a head and re-engineer the line.
These costs are capital expenditures, and almost all of them are fully expensible in the year of purchase under IRC Section 179 (currently a $1.16 M annual cap, well above what a single-line label-placement upgrade costs) or covered by bonus depreciation that, while tapering down, still applies. After-tax cost at a 21% corporate rate is approximately 79% of the gross figure — and for many meat brands organized as pass-through entities the effective rate is higher and the net cost lower still. The brand’s CFO sees this on the books at well below the headline number.
3.4.2 Continuing costs
These are incurred every year regardless of whether the EST# is on the front panel or the seam.
| Continuing item | Order of magnitude | Notes |
|---|---|---|
| Ink, ribbon, and consumables | $3,000–$10,000 per head per year | Same regardless of where the head prints — a head printing on the front panel uses the same ink as a head printing on the seam. |
| Print head maintenance contract | $1,500–$4,000 per head per year | Filter changes, ink reservoir refills, scheduled service visits. |
| Periodic head replacement | $1,000–$3,000 per replacement | CIJ ink modules are replaced roughly every 2–3 years as a routine service item. Full unit refresh is standard practice at 5–7 years. |
| Converter print-run setup amortization | Built into film price | The converter spreads plate setup cost across the run-length linear footage; the brand sees it as part of the per-foot film price, not as a separate line item. |
The key continuing-cost insight: there is no recurring premium for printing the EST# on the front panel rather than the seam. The print head runs the same hours, uses the same ink, draws the same maintenance contract, and gets refreshed on the same 5–7 year cycle no matter where its nozzle is pointed. The decision is essentially a one-time engineering decision; the operating cost is identical.
3.4.3 Print heads have to be swapped out anyway
This is the part that flips the framing. Industrial CIJ and TIJ print heads have a finite service life. On a meat-packaging line running 16 hours a day, the routine refresh interval is approximately 5–7 years for a full unit replacement, with ink-module replacements every 2–3 years inside that cycle. A USDA Large plant with 6–10 packaging lines is replacing or refreshing variable-data print heads on an essentially continuous rolling basis — one or two heads per year as a normal operating activity.
What this means in practice: the question is not whether to spend the capital, it is whether the next scheduled head replacement is mounted in a position that prints the EST# on the front panel or the seam. Re-positioning at the next refresh cycle adds engineering hours but very little incremental hardware cost. A brand that committed today to “the EST# goes on the front panel at the next print-head replacement on every line” would absorb the change inside its existing capital cycle over 5–7 years, with most of the per-line cost being engineering and validation rather than new equipment.
3.4.4 How this stacks against typical annual line capex
For context, a single packaging line in a USDA Large meat plant carries annual capital spend in the range of $300,000–$1.5 million depending on age, throughput, and ongoing automation projects. Routine line capex items include:
- New form-fill-seal or thermoform machine: $400,000–$1,200,000 (refreshed every 10–15 years)
- Weigh-price labeler: $80,000–$150,000 (every 7–10 years)
- Metal detector / X-ray inspection: $25,000–$80,000 (every 5–7 years)
- Vision inspection / camera system: $40,000–$120,000 (every 5–7 years)
- Conveyor upgrades and sanitation system improvements: $50,000–$200,000 per year, recurring
- Variable-data print head refresh: $15,000–$35,000 per head, multiple heads per line
The one-time $20,000–$110,000 redesign-and-reposition outlay to move the EST# to the front panel is 5–10% of a single year of routine line capital expenditure on a normal line, and is roughly equivalent in size to one routine print-head refresh that the brand was already going to make anyway.
3.4.5 Per-package cost
If a brand simply absorbs the full one-time outlay against a single year’s production:
- A mid-volume USDA Large plant produces roughly 50 million packages per year per line, sometimes more.
- A $75,000 one-time outlay (midpoint of the redesign + reposition range) spread across one year of one line: $0.0015 per package, or about 15 hundredths of a cent.
- Amortized across the typical 5–7 year refresh cycle of the affected equipment: under half a millicent per package.
The per-package cost is, by any reasonable interpretation, immaterial. It is a rounding error on the cost of the meat itself, the labor, the refrigeration, and the freight.
3.4.6 Sources we recommend verifying before final publication
The order-of-magnitude reasoning above is robust and is the load-bearing argument. The specific numbers should be checked against named industry sources before this paper is published under the FAT byline:
- Flexographic Technical Association — plate setup cost per color
- Videojet / Domino / Markem-Imaje — public CIJ print head capital and service-cycle data
- Packaging Strategies and Smithers PIRA — converter run-length norms for U.S. retail meat film
- USDA AMS — plant capital expenditure surveys, where published
The numbers may move within the ranges given, but the structural finding — one-time outlay in the low five figures per line, no recurring premium, swappable at the next routine print-head refresh, immaterial per package — does not move with them.
4. Why processors choose the cheaper path
Four reasons, in declining order of explanatory power:
The redesign cost is highly visible; the consumer cost is invisible. A redesign costs the brand $25,000 in plate work today, allocated to a single P&L line. The cost to consumers of not being able to see the EST# at the case is borne by the consumer, not the brand. There is no internal mechanism that surfaces the consumer-side cost in the redesign decision.
Pre-market label approval is one-and-done. Once a label clears LSDS review, the brand has no incentive to revisit the placement. FSIS does not re-review approved labels for consumer-side visibility unless a complaint or recall triggers a look.
Field enforcement is structured around the plant, not the consumer. In-plant FSIS inspectors verify that the label is being applied. The consumer-side question — can a shopper in the refrigerated case actually read the number — is functionally outside the inspector’s beat.
Front-panel real estate is contested. The brand mark, photography, claims, price sticker, weight, sell-by date, and allergen statement all compete for the front panel. The EST# is the lowest-priority required element from the brand’s perspective; it loses every internal real-estate fight.
A fifth factor, not the processor’s responsibility, completes the picture:
- Retailer-applied price stickers regularly cover the EST# even when the original processor’s placement was compliant. This is a retail practice, not a federal-meat-inspection violation, but it produces the same consumer-facing outcome: the disclosure is unavailable at the case.
- Retailer-applied price stickers regularly cover the EST# even when the original processor’s placement was compliant. This is a retail practice, not a federal-meat-inspection violation, but it produces the same consumer-facing outcome: the disclosure is unavailable at the case.
Convenience vs. Strategy
The mechanical explanation accounts for how seam placement became common. It does not explain why the industry has not moved away from it. The cost analysis above shows that front-panel placement is achievable for approximately 0.15 cents per package — an outlay that processors routinely absorb for cosmetic label changes.
The more complete explanation involves what the EST. # does when a consumer can find it. A clearly visible establishment number is a direct lookup key into FSIS’s public enforcement record — recalls, humane handling violations, chemical residue detections, pathogen test failures. Processors have access to that same public record and know what it contains.
Seam placement is not neutral. It is the outcome that minimizes enforcement-record lookups at negligible cost to the processor. The fact that seam or edge placement converged across the industry — across competitors, across species, across packaging formats — without any regulatory requirement to do so is consistent with a shared strategic interest rather than independent technical necessity.
FAT treats inaccessible EST. # placement as a disclosure failure regardless of whether the cause is equipment positioning or deliberate design. The effect on the consumer is the same.
5. FAT’s editorial position
The regulatory text already states the standard: prominent, conspicuous, legible, likely to be read by the ordinary individual under customary conditions of purchase. The flexibility in 9 CFR 317.2(c)(2)(ii) is a permissive option for the processor, not a license to bury required disclosure on a seam or under a glued flap.
The mechanical analysis in Section 3 demonstrates that compliance is not technically difficult. It is operationally costly in a one-time, redesign-front-loaded way, and the per-package amortized cost falls into a range that does not justify the consumer-facing disclosure failure.
FAT’s position:
Record the placement failure as a disclosure failure. Both the FAT App and the FAT website score “establishment number present but not legibly accessible” as missing disclosure under our 16-Category framework. This is consistent with the regulatory text and with consumer experience.
Distinguish processor-side from retailer-side failures. When a store-applied price sticker covers a compliant manufacturer label, the failure is the retailer’s. FAT will surface this distinction in scoring rather than collapsing both into a single processor demerit.
Recommend FSIS clarification. A targeted Federal Register notice clarifying that 317.2(c)(2)(ii) “elsewhere on the labeling” placements remain subject to the 317.2(b) “prominent and conspicuous” standard — and that “prominent and conspicuous” includes a presumption that the consumer should not need to break the seal, rotate the package, or look under a glued flap to read required disclosure — would close the gap without re-opening the underlying placement rule.
6. Open questions for the next paper in this series
- What is the actual NR / MOI rate for label-placement and legibility issues (Section 2)?
- How does the in-store retail handling — store-applied stickers, fold-over of vacuum-pack pouches, repositioning by stock clerks — degrade pre-market-approved label visibility?
- What is the actual industry-quoted cost of a compliant redesign, with named vendor sources?
- How do other USDA-regulated disclosures (Country of Origin labeling, allergen statements) compare on placement enforcement?
Sources cited
- 9 CFR 317.2 — Meat label requirements (definition, placement, conspicuousness, type size)
- 9 CFR 381.96 — Poultry parallel
- 9 CFR 412.1 / 412.2 — Label approval procedures
- FSIS Compliance Guidelines on Labeling Compliance
- FSIS Inspection Task data (NRs / MOIs), retrieved monthly via the FAT site’s FSIS automation pipeline
This paper was drafted on 2026-05-18 by Dirk Adams with the assistance of AI. The mechanical and cost figures in Section 3 are defensible estimates pending industry-source verification; the regulatory framework in Section 1 reflects the current text of 9 CFR; the citation count in Section 2 will be populated from the FAT site’s NR/MOI database before publication. This is the first paper in a planned FAT Research series on operational drivers of meat label disclosure failures.
