45.5%
of US FDA food recall events in 2024 were caused by label errors
State of Packaging Artwork Management 2026
The latest research on recalls, regulation, complexity and AI — and what they mean for the teams managing packaging artwork.
Executive summary
Packaging artwork has never had to carry more: more regulatory information, more variants, more markets, more digital data. The teams managing it have never been under more pressure.
This report brings together the latest third-party research on the forces reshaping packaging artwork management — recalls, regulation, SKU complexity, process and AI — plus Cway's own benchmark data.
The conclusion: the volume and risk of artwork change are rising faster than most artwork processes were designed to handle. Companies that treat artwork as a controlled, data-driven process, rather than a design task with email approvals, will absorb the coming regulatory waves with less cost and less risk.
New to the topic? Start with our explainer on what artwork management is.
Five numbers that define 2026
- 45.5% of US FDA food recall events in 2024 were caused by label errors, most of them undeclared allergens (Loftware analysis of FDA data).
- 12 August 2028 is the earliest date for PPWR harmonised labelling — the start of an EU-wide artwork update wave across every packaging SKU.
- End of 2027 is the GS1 Sunrise 2027 target for retail point-of-sale systems to scan 2D barcodes.
- From 70% to 18%: the share of packaging executives whose function had no generative AI activity fell this far in one year (McKinsey, 2024–2025).
- 24 days and 6.2 revision rounds is the average FMCG artwork approval cycle (Cway Packaging Artwork Approval Benchmark 2026).
Together they describe one trend: more mandatory artwork change, more risk per error, and processes that were not designed for either.
12 Aug 2028
earliest date for PPWR harmonised labelling across the EU
70% → 18%
packaging executives with no gen-AI activity in their function, 2024 → 2025
1. Recalls: the label is now the biggest single risk
Most recalls linked to packaging are not manufacturing failures. The product was made correctly; the label didn't match it.
United States
- In 2024, label errors were the leading cause of FDA food recalls: 192 of 422 recall events, or 45.5%. That was almost three times the second cause, listeria. 83.85% of those label-error recalls involved undeclared allergens. (Loftware analysis of the FDA Enforcement Report database, via New Food Magazine, Jan 2025)
- In 2025, US food and drink recalls reached a nine-year high, according to Sedgwick's 2026 Recall Index: 571 FDA recalls (+15.4% on 2024) affecting 138.5 million units. Undeclared allergens accounted for about 16 million recalled units. (Sedgwick via Just Food / Just Drinks, Feb 2026)
- In Q1 2026, undeclared allergens caused 57 of 140 FDA food recalls, the highest Q1 total in eight years. (TraceGains, US Product Safety and Recall Index, Jun 2026)
- Brand size is no protection. In late 2024, Frito-Lay recalled a limited batch of Lay's Classic chips over undeclared milk, and the FDA upgraded it to Class I, its most serious risk level, in January 2025. (ABC News, Jan 2025)
Europe and Australia
- EU food and feed safety alerts exchanged through the Alert and Cooperation Network rose 11% to 10,490 in 2025. RASFF notifications rose 2% to 5,344. (European Commission annual report, via Brussels Times)
- In Australia, undeclared allergens were again the leading cause of food recalls in 2025, at 38% of all recalls. (Food Standards Australia New Zealand)
Pharma
- A ten-year peer-reviewed analysis of FDA drug recalls (2012–2023) attributes about 19% to labelling and packaging issues. (cited by InformaIT)
- Industry estimates attribute 35–40% of pharmaceutical recalls to pack labelling errors and omissions. (Freyr Solutions)
The cost of a recall
- The average direct cost of a US food recall is about $10 million, and 5% of companies surveyed reported recalls costing over $100 million. (GMA / FMI)
- About 15% of consumers say they would never buy a recalled product again, and 21% would avoid the manufacturer's other products. (Harris Interactive, via Rentokil)

What it means for artwork teams: the artwork process — approved content, version control and comparison against the source of truth — is now a frontline food-safety control, not just a design workflow.
2. Complexity: more SKUs, more variants, more cost
Artwork volume no longer tracks the number of products. It tracks products × markets × languages × channels × regulatory changes.
- The global packaging market is worth about $1.2 trillion in 2025 and is forecast to reach about $1.52 trillion by 2030, at roughly 3.5–3.8% annual growth. (Smithers)
- Digitally printed packaging and labels are expected to grow from $22.0 billion in 2025 to $36.9 billion in 2030, a 10.9% CAGR. More short runs, versioning and personalisation mean more artwork variants. (Smithers, The Future of Digital Print for Packaging to 2030)
- E-commerce packaging was worth about $81.6 billion in 2025 and is forecast to exceed $100 billion by 2028. Each channel needs its own pack images and content. (Smithers × Easyfairs)
Complexity costs money
- SKUs with few shared components cost on average 40.8% more in total cost, and 105% more in packaging materials, than SKUs built from common components. (MIT Center for Transportation & Logistics, 2023)
- McKinsey estimates that managing complexity well can add 1–4 percentage points of net revenue and 3–6 points of margin.
- SKU rationalisation pays, but innovation keeps adding SKUs. L.E.K. Consulting estimates SKU reductions added 65–90 bps to brand owners' gross margins versus 2019. Brands that are adding SKUs cite product innovation, sustainable packaging and new pack sizes as the main drivers. (L.E.K. Annual Packaging Study)
What it means for artwork teams: processes built for a few hundred artworks a year break at a few thousand.
3. Process: where time is lost
Data from the Cway Packaging Artwork Approval Benchmark 2026:
- The average FMCG artwork approval cycle takes 24 calendar days.
- The average project involves 6.8 stakeholders and 6.2 revision rounds.
- 41% of teams still rely primarily on email-based approvals.
- Email-based workflows need 8.4 revision rounds on average, versus 4.1 in structured workflows.
- Each additional revision round adds about 2.3 days.
- 1 in 3 packaging projects misses its intended launch date.
Why manual review struggles
- Human proofreading has hard limits. Proofreaders catch around 90% of non-word typos but only about 70% of "real-word" errors (such as "1 mg" vs "10 mg"), and far fewer omissions. Omissions are exactly the error type behind undeclared-allergen recalls. (R. Panko, human error research)
- Errors enter late. Most artwork errors are introduced in late revision rounds, through a claim change, reformulation or translation update, not when the label is first designed.
4. Regulation: a decade of mandatory artwork change
Several regulations now create portfolio-wide, deadline-driven artwork updates, not one-off changes.
| Regulation | Region | Key dates for artwork | What changes on pack |
|---|---|---|---|
| PPWR (Regulation (EU) 2025/40) | EU | In force Feb 2025; applies from 12 Aug 2026; harmonised material/sorting labels from 12 Aug 2028 at the earliest (or 24 months after implementing acts); reusable packaging labels and QR codes from Feb 2029; recycled-content targets from 2030 | Harmonised pictograms replacing national schemes; material composition; sorting and reuse information; digital data carriers |
| FDA “healthy” claim rule | US | Compliance date 25 Feb 2028 | Products using “healthy” must meet the new criteria, or the claim comes off the pack |
| FDA front-of-package (FOP) nutrition labelling | US | Proposed Jan 2025; compliance proposed at 3 years after the final rule for companies with ≥$10M food sales (4 years for smaller ones) | A new “Nutrition Info” box on the front panel of most packaged foods |
| FDA uniform compliance date | US | 1 Jan 2028 for food labelling rules published 2025–2026 | Bundled label updates across ranges |
| GS1 Sunrise 2027 | Global (industry initiative, retailer-driven) | Retail POS able to scan 2D codes by end of 2027 | 2D codes (GS1 DataMatrix / QR with GS1 Digital Link) added next to the linear barcode, with new placement, sizing and print-quality checks |
| US state EPR laws (e.g. California SB 54, plus Oregon, Colorado, Washington, Maryland, Minnesota, Maine) | US | Phased, state by state | Packaging data, recyclability and environmental claims (FTC Green Guides, CA SB 343) |
Sources: TraceOne, FoodNavigator, Coolset, Paxly; Mintz, Hogan Lovells, Eurofins, FDA; GS1 / Videojet / Keyence.

What it means for artwork teams
- The work is cumulative. Between 2026 and 2030, a European or transatlantic brand may need to touch most of its artwork more than once, for PPWR labels, 2D codes, claims and FOP nutrition.
- Dates move. The PPWR labelling date depends on implementing acts, and the FDA FOP rule isn't final. Teams need a living regulatory calendar connected to their artwork pipeline.
- Bundle changes. Combining several mandatory changes in one artwork revision cuts repro and print costs, but only if you can see which SKUs are affected by which change.
Regulatory calendar 2026: PPWR applies
12 August 2026 — the EU Packaging and Packaging Waste Regulation (Regulation (EU) 2025/40) starts to apply across all member states.
- In force since February 2025
- Sets the framework for harmonised labels, material information and digital data carriers
- Every EU packaging SKU is in scope
Slide through to see what follows, year by year.
2027: 2D barcodes at retail checkout
End of 2027 — GS1 Sunrise 2027 target for retail point-of-sale systems to scan 2D codes.
- GS1 DataMatrix or QR with GS1 Digital Link added next to the linear barcode
- New placement, sizing and print-quality checks on artwork
- Industry initiative, driven by retailers
2028: the busiest artwork year
- 1 Jan 2028 — FDA uniform compliance date for food labelling rules published 2025–2026
- 25 Feb 2028 — FDA “healthy” claim rule: meet the new criteria or take the claim off the pack
- 12 Aug 2028 (earliest) — PPWR harmonised material and sorting labels replace national schemes
For transatlantic brands, three deadlines land on the same portfolio within eight months.
2029–2030 and beyond
- Feb 2029 — PPWR labels and QR codes for reusable packaging
- 2030 — PPWR recycled-content targets
- ~3 years after final rule — FDA front-of-package “Nutrition Info” box (4 years for companies under $10M food sales)
- Phased, state by state — US EPR laws (California SB 54, Oregon, Colorado, Washington, Maryland, Minnesota, Maine)
5. AI: from interest to early adoption
- In McKinsey's survey of packaging executives, the share whose function had not developed any generative AI capability fell from 70% in 2024 to 18% in 2025. In 2025, 13% had launched a gen-AI function, 33% were developing one and 38% were considering it. (McKinsey, presented at interpack 2026, via Baking Business)
- In the 2025 survey of 110 senior packaging leaders, 82% said their function had launched, was developing or was considering gen-AI solutions. 56% reported machine learning or gen-AI use in commercial excellence, versus 43% in procurement and 37% in supply chain. The main barriers were IP and privacy concerns and unclear value cases. (McKinsey, via Packaging Gateway)
- An earlier McKinsey survey of more than 200 paper and packaging executives found 95% believed their companies should invest in gen AI, and 77% had moderate to strong intentions to use it. (via Packaging Dive)
- 73% of respondents to Esko's Packaging Trends 2025 Survey named automation, AI and machine learning as the biggest impact on the industry in 2025. (Esko)
- On the marketing side, McKinsey estimates gen AI could lift marketing productivity by 5–15% of total marketing spend, about $463 billion a year. Yet in 2024 only 4% of consumer CMOs said they were scaling gen-AI use cases. (McKinsey)
Where AI helps artwork management, and where it doesn't (yet)
| Strong today |
Still needs deterministic checks and human sign-off |
|---|---|
| Flagging likely issues: missing elements, inconsistencies across variants, layout and legibility | Exact text comparison against approved copy |
| Running simple, user-defined rules at scale (e.g. allergen emphasis, mandatory statements present) | Barcode, 2D code and Braille verification with documented evidence |
| Finding and reusing past artwork content and decisions | Final regulatory, legal and brand judgement |
| Generating variants and digital pack imagery | Audit trails for regulated categories |
What it means for artwork teams: AI's value depends on clean inputs — approved content, structured product data and a single source of truth for files and decisions. Teams still running artwork through email and shared drives will struggle to use AI safely.
6. People: the artwork role is becoming a process-owner role
Cway analysis of around 130 artwork, packaging and brand job postings published in 2026 in Europe and North America.
- Hub models are spreading. Large groups increasingly run artwork from shared-service hubs, for example in Lisbon, Kraków, Warsaw and Wrocław.
- The scope is widening from print artwork to CGI pack images, e-commerce content, regulatory change programmes and stock-switch planning (SLOB and write-off avoidance).
- Employers measure outcomes: Right First Time, OTIF, SLA adherence, cycle time, quality incidents, revision rounds.
- Senior roles own the system and the process: selecting artwork management systems, running RCA/CAPA, leading teams and managing artwork budgets.
→ Full breakdown in Inside the Role of an Artwork Manager: Responsibilities and Skills.
7. What leading teams do differently: 7 recommendations
- Build one source of truth for approved content, files and decisions before automating anything.
- Map regulation to SKUs. Keep a living calendar of PPWR, FDA and GS1 changes, and know which artworks each one touches.
- Bundle mandatory changes into single revisions to cut repro and print costs.
- Verify changes, don't re-read everything. Use automated comparison against the previous approved version and the approved copy at every round.
- Make approvals predictable: defined reviewers, scopes and deadlines, and fewer, earlier stakeholder inputs.
- Prepare data for AI first: structured product data, templates and clean file libraries.
- Measure the process. Track cycle time, revision rounds, Right First Time and errors caught per stage.

Methodology and data notes
- This report combines publicly available third-party research (McKinsey, Smithers, Sedgwick, Loftware, TraceGains, L.E.K. Consulting, MIT CTL, GMA/FMI, regulators and industry bodies) with Cway's own research: the Packaging Artwork Approval Benchmark 2026 and an analysis of job postings.
- Figures are reported as published by their sources. They are not always directly comparable, because sources use different methods and definitions, for example recall events vs recalled units, or FDA vs FDA + USDA.
- Regulatory dates reflect information available in November 2026. Some dates depend on implementing acts or final rules and may change. This report is not legal advice.
- Market size of artwork management software: published estimates vary widely between research firms, from roughly $0.6 billion to several billion dollars, depending on definitions. We therefore don't quote a single figure.
Ready for the next regulatory wave?
See how your artwork process holds up against PPWR, GS1 Sunrise 2027 and new FDA labelling rules —
and where a structured, data-driven workflow would cut risk and revision rounds.
Talk to an artwork expert
FAQ
Packaging artwork management is the controlled process of creating, reviewing, approving, versioning and archiving the artwork printed on packs and labels. It connects approved content (ingredients, allergens, claims, regulatory text, barcodes) with design files and sign-offs, so that what is printed matches the product and the rules in every market.
In 2024, label errors caused 45.5% of US FDA food recall events (192 of 422), and 83.85% of those involved undeclared allergens. In these cases the product was made correctly, but the label did not match it. That makes approved content, version control and comparison against the source of truth a frontline food-safety control.
The EU Packaging and Packaging Waste Regulation (Regulation (EU) 2025/40) applies from 12 August 2026. Harmonised material and sorting labels follow from 12 August 2028 at the earliest (or 24 months after the implementing acts), reusable packaging labels and QR codes from February 2029, and recycled-content targets from 2030. Because the labelling date depends on implementing acts, teams should keep a living regulatory calendar.
GS1 Sunrise 2027 is a retailer-driven industry initiative for point-of-sale systems to scan 2D codes by the end of 2027. For artwork teams it means adding a GS1 DataMatrix or QR code with GS1 Digital Link next to the linear barcode, with new placement, sizing and print-quality checks.
Key dates include 1 January 2028, the uniform compliance date for food labelling rules published in 2025–2026, and 25 February 2028, the compliance date for the new “healthy” claim rule. The proposed front-of-package “Nutrition Info” box would apply 3 years after the final rule for companies with at least $10M in food sales (4 years for smaller ones). State EPR laws add packaging data and recyclability-claim requirements.
According to the Cway Packaging Artwork Approval Benchmark 2026, the average FMCG artwork approval cycle takes 24 calendar days, with 6.8 stakeholders and 6.2 revision rounds. Email-based workflows need 8.4 revision rounds on average versus 4.1 in structured workflows, and each extra round adds about 2.3 days.
AI is strong at flagging likely issues (missing elements, inconsistencies across variants, legibility), running user-defined rules at scale, finding past content and decisions, and generating variants and digital pack imagery. Exact text comparison, barcode and Braille verification, final regulatory judgement and audit trails still need deterministic checks and human sign-off. AI also depends on clean inputs: approved content, structured product data and a single source of truth.
Build one source of truth for approved content and files, map each regulation to the SKUs it affects, bundle mandatory changes into single revisions, use automated comparison instead of re-reading everything, make approvals predictable, prepare data for AI, and measure cycle time, revision rounds and Right First Time.
Sources
- Label errors dominate 2024 US food recalls — New Food Magazine (Loftware)
- US food & drink recalls hit nine-year high — Just Food (Sedgwick)
- Sedgwick: Five industries saw higher US recall volumes in 2025 — Claims Journal
- FDA food recalls hit 6-year high — TraceGains
- FDA upgrades Lay's recall — ABC News
- EU food alerts soar 11% — Brussels Times
- 2025 food recall statistics — FSANZ
- Why pharma artwork errors happen at the end of the process — InformaIT
- Right First Time artwork capabilities — Freyr Solutions
- Costs of food recalls — Food Safety Magazine
- The cost of product recalls — Rentokil
- What August 12, 2026 means for PPWR — TraceOne
- PPWR timeline — FoodNavigator
- PPWR compliance deadlines explained — Coolset
- FDA continues push to improve food labeling — Mintz
- Front-of-package nutrition labeling — FDA
- FDA general food labeling compliance program update — Eurofins
- GS1 Sunrise 2027 — Videojet
- Smithers: tariffs to reshape $1.5 trillion packaging market
- Smithers: digital print for packaging to 2030 — Screen Print India
- Four trends redefine the future of packaging — Packaging World (Smithers × Easyfairs)
- Quantifying packaging material complexity — MIT CTL
- Managing complexity in consumer goods — McKinsey
- What happened to SKU proliferation? — L.E.K. Consulting
- Sustainability in packaging 2025 — McKinsey
- Packaging industry growth tied to generative AI — Baking Business (McKinsey at interpack 2026)
- AI in packaging: where will it have the biggest impact? — Packaging Gateway
- 5 ways AI is shaping packaging today — Packaging Dive
- AI and packaging — Esko
- How generative AI can boost consumer marketing — McKinsey
- McKinsey: CMOs and gen AI scaling — Retail Times
- Human error research — R. Panko (arXiv)
- The Packaging Artwork Approval Benchmark 2026 — Cway