What GS1 and Related Barcode Standards Actually Define, and What to Watch for in Real Operations - Organizing GTIN, AI, GS1-128, and GS1 DataMatrix

In barcode discussions, it is very common for requests like we want JAN, we want GS1-128, and we want QR to get mixed together as if they were all the same kind of decision.

In reality, there are several things that should be separated first:

• what you want to identify
• what additional attribute information you want to carry
• which symbol should represent it
• how the upper system should interpret the scanned result afterward

GS1 is the standard that connects those four layers. In other words, GS1 is not only about “how numbering works.” It is also a contract that includes the meaning of the data, the barcode carrier, placement and quality, and even how the scanned result should be handled later.¹

The discussion below is organized around GS1, GS1 Japan, and Ministry of Health, Labour and Welfare materials that could be confirmed as of April 2026.

Contents

1. Table of Contents
2. 1. The short answer
3. 2. The layers standardized by GS1
   • 2.1 Identification keys
   • 2.2 AI defines the meaning of attribute data
   • 2.3 Barcode symbols as data carriers
   • 2.4 Placement, size, quality, and HRI
4. 3. The practical use cases you should separate first
   • 3.1 Consumer-facing POS
   • 3.2 Logistics, shipping / receiving, and SCM labels
   • 3.3 Traceability and regulated fields
5. 4. Representative rules defined by GS1
   • 4.1 GTIN is not a number you can freely recycle
   • 4.2 AI has fixed-length and variable-length forms
   • 4.3 FNC1 is not a visible character
   • 4.4 It is safer to treat scanner output format as part of the standard too
   • 4.5 Putting 2D on the label does not mean you can replace everything immediately
6. 5. Points that often cause trouble in real operations
   • 5.1 Overwriting GTIN in the product master without version discipline
   • 5.2 Turning the human-readable bracketed format into the input specification
   • 5.3 Do not accept “it scanned once” as the whole acceptance test
   • 5.4 Do not judge print quality from how it looks on the monitor
   • 5.5 Do not arrange multiple barcodes casually
   • 5.6 Failing to confirm which AI are actually used in the field
7. 6. Cases where regulation or industry requirements are strong
   • 6.1 Retail 2D migration
   • 6.2 Medical devices and similar fields in Japan
8. 7. A practical rollout order
9. 8. Summary
10. 9. References

Table of Contents

1. The short answer
2. The layers standardized by GS1
3. The practical use cases you should separate first
4. Representative rules defined by GS1
5. Points that often cause trouble in real operations
6. Cases where regulation or industry requirements are strong
7. A practical rollout order
8. Summary
9. References

1. The short answer

When working with GS1 in practice, these are the five points worth fixing in your head first.

1. GTIN and barcode type are different things.
   GTIN is the product identifier, while EAN/UPC, ITF, GS1-128, GS1 DataMatrix, and GS1 QR Code are ways to represent it. If you mix those up, the design starts to collapse.²³

2. GTIN alone is often not enough.
   If you need to handle batch number, expiration date, or serial number, you need GS1 Application Identifiers (AI) so those attributes can be carried separately.⁴

3. The right symbol differs between POS, logistics, and traceability.
   Consumer POS usually centers on GTIN, logistics often centers on GTIN-14 or SSCC, and traceability or regulated fields usually build around GTIN plus lot, expiry, and serial data.²⁵

4. Operational failures happen more often on the receiving side than in the print itself.
   If FNC1, variable-length AI boundaries, scanner settings, symbology identifiers, upper-system parsers, and product-master update rules are undefined, the code may scan but still not work in business use.⁴³

5. Retail 2D migration is advancing, but it is still a transition period.
   GS1’s retail 2D guidance assumes that until POS environments are broadly ready for 2D, a 2D code may still need to be printed alongside the linear barcode used at POS today.⁶⁷

So instead of starting from the question of how to generate a barcode, it is safer to start from who reads what, where, and for which business purpose.

2. The layers standardized by GS1

2.1 Identification keys

GS1 begins with identification keys that answer the question, “what exactly are we identifying?”

The 대표 example for products is GTIN. GTIN comes in GTIN-8 / 12 / 13 / 14 forms, and in Japan GTIN-13 and GTIN-8 are widely used as JAN codes. GTIN-13 is used for single items or the smallest trade unit, while GTIN-14 is used for grouped packaging such as cases and cartons.²

A representative logistics key is SSCC. Attaching SSCC to pallets or shipping cases and linking it to shipment data is a very GS1-like way to simplify receiving and inspection.⁵

The important point here is that GTIN tells you which product it is; it is not the same thing as lot or expiry information. Lot and expiry live in another layer.

2.2 AI defines the meaning of attribute data

Additional attributes attached to GTIN are defined through GS1 Application Identifiers (AI). An AI is a 2-digit to 4-digit number that defines both the meaning and the format of the data that follows it.⁴³

Representative examples include:

• (01) GTIN
• (00) SSCC
• (10) lot / batch number
• (17) expiration date
• (21) serial number

What matters in GS1 is not merely that a numeric string can be scanned, but that the same numeric string can be interpreted with the same meaning. AI makes it easier for the receiving side not to confuse multiple data elements even when they are packed into one barcode.⁴

2.3 Barcode symbols as data carriers

The next question is what will actually carry that data. In the GS1 world, that is treated as a separate layer as well.¹⁸

For cases that only represent GTIN, EAN/UPC (JAN symbols in Japan) and ITF are still widely used. If you also need expiration dates, lot numbers, or other attributes, then GS1-128, GS1 DataMatrix, GS1 QR Code, and similar options become candidates.³

A common misunderstanding here is that GS1-128 is not simply “Code 128 itself.” In the same way, GS1 DataMatrix is Data Matrix under GS1 rules, and GS1 QR Code is a GS1-defined subset of QR Code. They use FNC1 to mark the data as GS1 data and then arrange the content according to AI syntax.⁸³

So “I can generate Code 128” and “I can generate a correct GS1-128” do not mean the same thing.

2.4 Placement, size, quality, and HRI

The GS1 General Specifications cover not only identification keys and AI, but also symbol size, X dimension, quiet zones, placement, quality grading, and Human Readable Interpretation (HRI).¹

That means barcodes are not design elements that can simply be scaled freely up and down. Especially in logistics or high-speed POS contexts, print quality and placement directly affect read rate. Even GS1’s retail 2D guidance treats placement and the operation of multiple barcodes as directly related to POS throughput.⁶

3. The practical use cases you should separate first

3.1 Consumer-facing POS

At retail checkout, the first priority is reliable GTIN reading at speed. There is increasing interest in carrying more attributes, but today is still a transition period, and in many cases 2D needs to be printed alongside linear codes rather than replacing them outright.⁶⁷

That is why “QR is the future, so we should make everything 2D” is risky as a decision rule. It is not enough to know whether the scanner can read it; you also need to confirm what the host POS can receive and what it can actually use.⁶⁷

3.2 Logistics, shipping / receiving, and SCM labels

In logistics, the focus is on identifying cases and pallets, receiving inspection, and matching with shipping data. In that context, combinations such as GTIN-14 or SSCC together with GS1-128 are still very strong in practice.²⁵

In other words, logistics is often less about “what is inside” than about “which physical unit is tied to which operational data.” The design tends to work better when the barcode is planned together with ASN or shipment-record data rather than as an isolated label.

3.3 Traceability and regulated fields

In medical, food, safety, and expiry-controlled areas, GTIN alone is often not enough. GTIN is combined with lot, expiration date, and serial number so that traceability can work at the item or batch level.⁴⁹

The GS1 DataMatrix guideline also explains that GS1 DataMatrix is primarily intended for open-system use, where trading partners are expected to read the same data with the same meaning.⁸

In these contexts, the real question is not only whether it can be printed, but whether the same meaning can flow across trading partners and regulators.

4. Representative rules defined by GS1

4.1 GTIN is not a number you can freely recycle

In GTIN assignment, one key question is whether the old number can still be kept. GS1 Japan presents three major principles for when a new GTIN is required after launching a new product or changing an existing one:¹⁰

• consumers or trading partners would want to distinguish it from the previous product
• regulation or liability disclosure requires separation
• an important change affects shipment, storage, receiving, or another part of the supply chain

In practice, teams often hesitate over things like size, quantity, packaging form, sales unit, and changes that affect labeling responsibility. If those are handled only as business-side judgment, the distortion appears later in EC, ordering, inventory, and master-data integration.

In Japan, GTIN reuse has in principle not been allowed since 2019.¹¹

That is a very important point. If GTIN is treated like an internal code that can be reused once it becomes free, it collides with historical product information and EC history. It is safer to think of GTIN as a long-lived key whose meaning persists over time, including product masters and externally published data.

4.2 AI has fixed-length and variable-length forms

With GS1 AI, some data elements are fixed length and others are variable length. When a non-fixed-length AI is followed by another element, separation with FNC1 is required. In the GS1 syntax dictionary, predefined-length AI are marked with *, indicating that FNC1 is not needed there.⁴

For example, the AI (10) for lot number is variable length. If another AI follows it, the receiving side must still be able to determine where the (10) value ends. If this is implemented loosely, you can produce a barcode that looks plausible but cannot actually be parsed correctly.

In practice, placing fixed-length elements first and variable-length ones later reduces separators and tends to produce much more stable results.¹²

4.3 FNC1 is not a visible character

FNC1 does not mean “insert the text FNC1.” GS1 Japan’s basic guide explains it as a control character used in GS1 standard barcodes, both to indicate GS1 data at the start of the symbol and to act as a data separator where needed.³

Also, the way it is represented is not identical across barcode types. For example, GS1 QR Code does not treat mode indicators and data separators in exactly the same way as 1D symbols do.³

The practical point is: do not assume internal representation, scanner output, and HRI display are the same string.

4.4 It is safer to treat scanner output format as part of the standard too

One thing that is easy to miss in system integration is the symbology identifier added after decoding by the scanner. GS1 Japan documentation shows examples where GS1-128 is reported upward as ]C1, GS1 DataMatrix as ]d2, and GS1 QR Code as ]Q3.³

That is not printed literally in the barcode itself. It is information added by the scanner to tell the upper system what kind of symbol was read. So the receiving side should decide one of these approaches up front:

• receive the symbology identifier and branch processing on it
• suppress or transform it in the scanner, so that the application receives only already-normalized strings

If UI and database design are done before that is settled, changing the inspection terminal later can break every equality check.

4.5 Putting 2D on the label does not mean you can replace everything immediately

GS1’s retail 2D guidance positions 2D barcodes as the likely future center, but also states that this is still a transition phase and that a linear barcode still needs to be printed alongside 2D until POS 2D readiness is sufficiently broad.⁶

Also, even among 2D options, GS1 DataMatrix may be comparatively straightforward to read while QR / Data Matrix flows built around GS1 Digital Link depend much more on the readiness of updated camera scanners and host-side support.⁶⁷

“A scanner exists that can read it” and “our POS / WMS / ERP can use it as part of operations” are not the same thing.

5. Points that often cause trouble in real operations

5.1 Overwriting GTIN in the product master without version discipline

Changing the meaning of a GTIN in place, or assigning an old GTIN to a different product later, gradually breaks EC, ordering, inventory, analytics, reports, and database registration.¹⁰¹¹

In real operations, GTIN assignment, product-master maintenance, and label issuance are often owned by different people or teams. That is why it is safer to design not only the numbering rules, but also the request and history-management flow for changes.

5.2 Turning the human-readable bracketed format into the input specification

In GS1 explanatory materials and HRI, AI is often written with parentheses such as (01) or (17). That exists so people can visually separate the meaning more easily.⁶

On the other hand, scanner output and machine parsing use different forms that may involve FNC1 or symbology identifiers.
So in practice, it is safer to separate the display string from the machine-processing string.

5.3 Do not accept “it scanned once” as the whole acceptance test

GS1 also defines invalid AI combinations and cases where a particular AI must appear together with another element.⁴³

That is why “it scanned one time in the print test” is not enough. At minimum, it is better to check these separately:

• whether the symbol can be read as a symbol
• whether the expected string is received
• whether AI boundaries and lengths are correct
• whether the upper system interprets it according to the business rules
• whether it stays consistent with the product master, shipment data, and database

Scan success is not business success.

5.4 Do not judge print quality from how it looks on the monitor

The GS1 General Specifications define X dimension, quiet zones, and quality grading rules. In logistics, the required dimensions and whitespace also differ between GS1-128 and 2D symbols.¹

In real operation, read rates can drop sharply because of very ordinary things such as:

• label-printer resolution and thermal-transfer conditions
• bleed on cardboard or film
• gloss and reflection
• printing on curved surfaces
• casual downscaling during design adjustment
• layout changes that squeeze the quiet zone

So the real standard should not be what looks readable in Illustrator or PDF, but whether the actual printed object is readable at the actual distance and angle with the actual scanner.

5.5 Do not arrange multiple barcodes casually

The retail 2D guideline explicitly says that barcode placement rules affect POS throughput.⁶

In real work, trouble often appears in combinations such as:

• putting a marketing QR code too close to the POS JAN code
• placing linear and 2D codes too close together
• pushing expiry or lot HRI into a hard-to-read place
• using back, side, and hanging-tag placements where the scan direction is inconsistent

So with multiple codes, the real design question is not only “do they fit?” but which one should be read first and most reliably.

5.6 Failing to confirm which AI are actually used in the field

Even GS1’s retail 2D guidance notes that not every retailer supports every AI, and that available AI differ depending on the reading point and the use case.¹³

For example, the manufacturer may want to encode GTIN + expiry + lot, but if the receiving side uses only GTIN, that extra value exists only upstream. On the other hand, if expiry will be used in the field, then WMS, stocktaking, and disposal decisions all need to support it.

It is often easier to move a deployment forward when you confirm the data you can encode separately from the data people will actually use.

6. Cases where regulation or industry requirements are strong

6.1 Retail 2D migration

GS1 has published standardization and migration guidance for 2D barcodes at retail POS and is clearly moving toward a future where 2D plays a central role. But today is still a transition period, so keeping linear codes alongside 2D is often the practical choice until POS reading and host capability become broad enough.⁶⁷

So if you want to place 2D on B2C products, it is safer to confirm at least these separately:

• can the scanner read it
• can the host POS interpret it
• does the linear code still need to stay
• do HRI and placement actually work in checkout operations

6.2 Medical devices and similar fields in Japan

In Japanese medical devices, in-vitro diagnostics, and similar fields, notices already specify concrete operations such as product-code and manufacturing-identifier display by packaging unit, and use of GS1-128 or GS1 DataMatrix as the barcode symbology.⁹

The important point here is that GS1 is not merely a convenient private-sector rule.
In practice, laws, notices, industry guidelines, database registration, and hospital / distribution operations can all be layered on top of GS1 assumptions.

So in this kind of field, “it works inside our company” is not enough. The design has to use GS1 in a way that matches regulation, trading partners, and field operations.

7. A practical rollout order

If you want barcode introduction to be hard to break, a sequence like this is usually easier to manage:

1. Decide the business use case
   Is this for POS, inbound / outbound inspection, or traceability?

2. Decide the identification key
   GTIN, GTIN-14, or SSCC? If needed, include GLN and similar keys too.

3. Decide the attributes you actually need
   Lot, expiry, serial, weight, price, and so on, defined AI by AI.

4. Decide the symbol
   Which is natural here: EAN/UPC, ITF, GS1-128, GS1 DataMatrix, or GS1 QR Code?

5. Decide the scanner output specification
   Will you receive symbology identifiers? How will FNC1 be normalized? What will be handed to the host?

6. Decide the product-master and change rules
   Make new GTIN issuance, change, retirement, non-reuse, and history management explicit.

7. Test print on the actual object
   Use the actual label, actual packaging material, actual scanner, and actual distance.

8. Test the business flow, not just the scan
   Confirm not only “it can be read,” but “stocktaking, receiving inspection, shipping, and expiry handling all work.”

In other words, deciding the receiving-side specification before choosing label software is usually the faster path.

8. Summary

The most important first step with GS1 and related barcode standards is to
treat numbering, attributes, symbols, and post-scan processing as separate design layers.

• GTIN is the key that identifies the product
• AI defines the meaning and format of the data
• GS1-128 / GS1 DataMatrix / GS1 QR Code are representation methods
• quality, placement, FNC1, scanner output, and product-master operations all belong to the implementation

Once those are separated in the design, the barcode stops being “something that can merely be printed” and becomes a business interface that can be shared with trading partners and field operations.

If you keep those layers mixed together,
the barcode may scan, but inventory, receiving, and traceability still break.

So when you use GS1, it is more practical to start not from which code to print, but from who reads what, in which system, for which purpose.

9. References