GS1 Sunrise 2027 migration guide for CPG brands

What this guide deliberately does not cover

It does not explain how to build a code. There is no step-by-step on choosing a domain, assembling the URI, standing up a destination or generating the symbol, and nothing here about encoding settings or symbol sizing. That material is written up separately, in our guide to creating a GS1 QR code.

It also does not repeat the standards explainer. If you want the structure of a Digital Link address and what sits behind it, that is on the bilingual packaging page, which is written for the Canadian bilingual case rather than as a generic introduction. Terms are defined in the glossary.

What is left is the part nobody writes down. What order, what blocks what, and what is going to bite you.

Why sequencing is the whole problem

Every step in this transition inherits the output of the step before it. That sounds obvious until you notice how many of the steps are irreversible on a normal budget.

Product data flows into the code. The code goes into artwork. Artwork goes to a printer. Printed packaging goes into a print run that may sit in inventory for a long time. A mistake caught at the data stage costs an afternoon. The same mistake caught after the run costs the run.

So the useful question is not “how fast can we do this”. It is “what is the last responsible moment to decide each thing, and what am I unable to change afterwards”.

There are only two things in this process you cannot change later without real money. The physical packaging, and the web address printed on it. Everything else is software and can be revised. Protect those two, and the rest of the project is forgiving.

Step one: your product data, before anything else

This is the unglamorous prerequisite that determines whether the whole thing is cheap or awful, and it has nothing to do with barcodes.

A linear barcode carries a number to a checkout. Nobody sees it. Nobody checks it. If the net weight in your item records disagrees with the net weight on the panel, no shopper ever finds out, and the error can sit in your systems for a decade without consequence.

A 2D barcode changes the audience. The same identifier now leads a shopper with a phone to information about the product. Whatever you publish there is consumer-visible, in both official languages, sitting next to the physical package they are holding. Any disagreement between the two is now findable by anyone, including a competitor, a regulator or a journalist.

That is the real prerequisite. Not the symbol. The truthfulness of the data behind it.

What wrong data actually looks like

In practice, the problems are rarely dramatic. They are the accumulated small drift of systems that were never checked against each other.

• Identifiers assigned to products that were discontinued years ago, and identifiers quietly reused for something else.
• Net contents, package format or case configuration that differ between your ERP, your artwork files and what you send to retailers.
• Product descriptions that were written for an internal system, with abbreviations and codes that make no sense to a shopper.
• Ingredient and allergen statements that are current on the package but stale in the database, or the reverse.
• French copy that exists on the panel but was never captured anywhere structured, so the only true source is a PDF of the artwork.
• Several internal systems that each believe they are the source of truth, and a spreadsheet that is actually the source of truth.

None of that is unusual. Most brands have some of it. The point is that a 2D barcode does not create these errors. It surfaces them, all at once, in front of consumers.

Name an owner, because usually nobody owns it

Ask who is accountable for product master data at your company and you will often get a pause. Marketing owns the copy. Operations owns the ERP record. Quality owns the panel. Sales owns what goes to the retailer. The data itself belongs to everyone, which means it belongs to no one.

The single most useful deliverable in the early phase of this project is not a barcode. It is one named person who owns product data correctness, with the authority to say that a record is wrong and have it fixed.

If you do only one thing this year, do that one. It pays for itself whether or not you ever print a 2D code.

Step two: ask your retailers, and ask properly

The question that decides your timing is not whether your market is ready. It is what the point of sale roadmap looks like at the handful of accounts that actually carry your products. Those are different questions with different answers, and only the second one is answerable.

You are unlikely to get a useful answer by email to a general mailbox. Raise it with your category contact, ideally inside a meeting that already exists: a line review, a category review, a planning session. Ask more than one retailer, because the answers will not match, and the spread between them is itself the useful information.

Do not expect one contact to speak for the whole chain. Scanning is a store systems decision, packaging specs sit with a different team again, and the person who buys from you may genuinely not know.

The questions worth putting to a category manager

Take these to the meeting as they are written.

1. Can your lanes read a 2D barcode today? Both the fixed scanners and the handhelds, and in every store or only some formats.
2. If not, is there a schedule? Is it a hardware refresh, a software upgrade, or both, and does anybody own it internally.
3. If a pack carries both symbols, what happens at the lane? Which one gets scanned, has that been tested in your environment, and what does a cashier do if the wrong one is read.
4. Is there a supplier guideline or packaging spec for 2D codes? If yes, ask for the current version and how you will be told when it changes.
5. Are you running a pilot with any suppliers? If yes, what would it take to be in it. A retailer pilot is the cheapest possible way to learn what that retailer will eventually ask for.
6. Does adding a symbol to existing artwork trigger anything on your side? A new item submission, a re-approval, a packaging change notification, a photography refresh. This one catches people out.
7. Do you have requirements for what sits behind the code? Some retailers will have opinions about what a shopper sees when they scan a product in their store, particularly about anything promotional.
8. Does the data behind the code have to agree with what we already send you? In most cases it should anyway, but ask, because it tells you which internal system is about to become load bearing.
9. Who else at your end should be in this conversation? Almost always more than one team.
10. If we change nothing, does anything change for us? Scorecards, compliance programs, listing requirements. Ask directly and write the answer down.

How to read the answers

Listen for the difference between capability, intention and requirement. “Our scanners can read it” is capability. “We are upgrading next year” is intention. “You must supply it by this date” is a requirement, and it is the only one of the three that changes your plan.

If nobody is requiring anything, your timing is genuinely yours. That is not a reason to ignore the question. It is a reason to stop treating a vendor's calendar as your deadline.

Be equally careful with a warm yes. Category managers are helpful people and will often say something encouraging about a supplier initiative. That is not a commitment from store systems. If the answer matters to your artwork schedule, get it from the team that owns scanning.

Separately, your GS1 member organisation licenses identifiers in your territory, and it is worth asking them directly about current programs and guidance. GS1 Canada, GS1 US and the national organisations across Europe are separate bodies with their own programs and their own timing. They speak for the standards and for what they themselves publish. They do not speak for any retailer's upgrade schedule. What differs by market is covered on the Canada, United States and European Union readiness pages.

Step three: decide who controls the destination

The code you print is a web address. Somebody owns the domain in that address. Get this decision wrong and it is permanent for the life of the packaging, which is longer than most commercial relationships.

The mechanics are simple enough: the address points at something you can update, so the content can change without a reprint. Fine. The part that matters commercially is the word “you”.

Ask, before artwork:

• Whose domain is printed on the pack? If it is a platform's domain rather than yours, every future scan depends on that company continuing to exist and continuing to serve you.
• Who holds the registration and the DNS? Not “who manages it for us”. Who is the registrant on the record, and who can log in.
• What happens at the end of the contract? Can you take the mapping with you and point the same printed address at your own infrastructure. Get the answer in writing before the run, not during a renewal negotiation.
• What happens if the vendor is acquired or shuts down? Packaging outlives startups.
• Does the printed address read sensibly to a human? A shopper who can see your brand in the address has a chance of noticing when a sticker has been placed over the real code. That is a practical tampering concern on physical packaging, not a theoretical one. There is more on it in the FAQ.

None of this requires a particular vendor or a particular architecture. It requires that the address printed on a package you will be selling for years belongs to you.

The expensive version of this mistake is not subtle. A brand prints a vendor domain across a full packaging refresh, changes platform two years later, and discovers that every pack in the market points at an address they cannot control. The only fix is new packaging.

Step four: artwork, and the multilingual panel problem

Here is where the schedule actually lives. Not in the technology.

Packaging in a multilingual market is denser than in a single-language one, because it is doing the same job more than once. Canada is the clearest case: federal labelling rules require most prepackaged consumer products to carry core information in both English and French, and Quebec has its own French-language requirements on top of that. The EU version is larger again, since member states set their own language requirements. The mandatory content is therefore duplicated, the panel is already tight, and on small formats it is genuinely full. A single-language market like the US has no equivalent duplication, so the panel pressure is lower. Verify the specifics for your category with the relevant regulator, because they differ by product type.

Now add a symbol that needs clear space around it and cannot be shrunk past the point where it scans. On a large carton this is a minor layout exercise. On a small pouch, a sachet, a narrow bottle or a shrink sleeve, it is a real design constraint and it can force decisions about what moves or shrinks to make room.

Three practical consequences.

One more thing that is specific to this market. What the shopper sees after scanning is language-sensitive, and serving a French page to a French speaker is not an optional nicety here. That belongs in your plan from the start rather than as a phase two improvement, and it is covered in more depth on the bilingual packaging page.

Step five: pilot one line, properly

A pilot on a single product line will teach you more in a quarter than a year of catalogue-wide planning. Planning produces documents. A pilot produces facts, including the ones nobody predicted.

Choosing the line

Pick for learning, not for prestige.

• A line with an artwork revision already scheduled, so you are not paying for the cycle twice.
• A printer and substrate that are representative of the rest of your portfolio. Learning on your easiest pack teaches you nothing about your hardest one.
• Reasonable volume. Enough to be real, small enough that a mistake is affordable.
• Not the flagship. Not the SKU with the highest velocity and the most internal attention.
• Ideally a line sold into a retailer contact who will actually talk to you about it.

What the pilot has to prove

Write the success criteria down before you start, because otherwise the pilot ends with a vague good feeling and no decisions.

That last one is the real output. Most of the value of a pilot is an accurate number for how long this takes at your company.

Step six: print and substrate, the failure that arrives late

This is the surprise that costs the most, because it appears after the money is spent.

A 2D code that scans perfectly on a screen and in a PDF proof can fail on the actual pack. The symbol is fine. The reproduction is not.

Common ways it goes wrong:

• Matte, soft touch and textured finishes scatter light and can kill the contrast a camera needs.
• Flexible film stretches and distorts on the line, and a symbol that is geometrically perfect in the file is not perfect on the pouch.
• Shrink sleeves deform by design. Whatever you print gets curved and compressed, unevenly.
• Foil, metallic inks and high gloss varnish create glare that reads as noise.
• Dark or coloured substrates reduce contrast, and unconventional colour combinations that look beautiful in the brand guide can be unreadable to a scanner.
• Absorbent stock and ink spread blur edges, especially at smaller sizes and on fast presses.
• Curved and small surfaces mean the code is never flat and never fully in view at once.

The fix is process, not cleverness. Involve the printer before artwork is final and ask what they have actually printed and scanned on this substrate, not what they believe is possible. Then test the printed article, not the file.

A workable test: pull samples from the actual press run, not a digital proof. Scan them with several phones including an older one, in poor lighting, at the distance a shopper would really hold the pack, on a curved or crumpled sample as well as a flat one. If you can borrow a retail handheld scanner, use it too. Do this while the run is still stoppable.

Screen tests prove nothing about packaging. That is the whole lesson.

What blocks what, and what can run in parallel

Most of this project is not a chain. Treating it like one is how six months disappear.

Four things can start this week and none of them cost much: cleaning data, naming an owner, asking retailers, and settling the domain question. They are also the four that everything else depends on.

One nuance on the content behind the code. It does not block printing, because you can improve it later. But something correct and bilingual has to be live before the first pack reaches a shelf. Plan for that date, not the print date.

The mistakes that cost the most

Ranked roughly by how much they hurt.

1. Printing a web address you do not control. Permanent for the life of the packaging.
2. Printing before the data is right. Every downstream system inherits the error, and now consumers can see it.
3. Skipping press-side scan testing. The cheapest possible test, caught at the most expensive possible moment.
4. Starting with the whole catalogue. Multiplying an unvalidated process by four hundred SKUs.
5. Treating it as a packaging project. It is a data project with a packaging deliverable. Run by packaging alone, it produces a symbol and no substance.
6. Two unrelated QR codes on one pack. A marketing code and a product code competing for the same shopper, and usually for the same square centimetre.
7. Accepting a vendor's deadline as an industry deadline. They are not the same thing, and the difference is worth a lot of money.

What happens if you do nothing

Nothing happens. That is the honest answer and it deserves to be stated plainly.

Sunrise 2027 is a voluntary industry readiness target aimed at retailers and the companies that supply their scanning systems. It is not law, not a regulation and not a mandate on brands. There is no end date for the linear barcode, no requirement to remove it from packaging, and no penalty that attaches to a brand for not having a 2D code. A company that does nothing at all is not out of compliance with anything.

What you lose by doing nothing is not compliance standing. It is lead time and optionality.

If a retailer does eventually ask, there is a real gap between a brand that has clean data, an owned domain and one line already piloted, and a brand starting from zero. The gap is not the technology, which is fast. It is that artwork cycles, data cleanup and internal sign-off are slow, and slower still when they all start at once and nothing has been tested. How long that takes depends entirely on your own artwork and approval cycles, which you can measure and we cannot.

There is also a quieter cost. Every artwork revision you run between now and then is an opportunity you either use or spend. Packaging changes are expensive precisely because they are infrequent, so the marginal cost of adding a symbol during a revision you were doing anyway is small, and the cost of opening a revision solely for a symbol is not.

So the defensible position for a brand with no retailer pressure is not “do nothing”. It is “do the cheap, reversible parts now, and let the expensive irreversible part ride the next revision”. Clean the data. Own the domain. Ask the questions. Then wait, informed, instead of waiting blind.