QR codes are the barcode most people can name, but they are only one member of a family of two-dimensional codes. Data Matrix, Aztec, PDF417 and Micro QR all encode data in a printed pattern that a camera or scanner can read, and each was designed around a different set of problems: tiny parts, crumpled tickets, identity cards, or very small labels. This article explains what each code looks like, what it is genuinely good at, and why QR still wins for most everyday phone-camera use.
What "2D code" actually means
Traditional barcodes, like the ones on grocery packaging, store data as a row of parallel lines. They hold only a small amount of information, typically a short numeric identifier, because the data runs in one direction only. Two-dimensional codes store data in both directions, using the full area of the symbol. That single change makes a huge difference in capacity: instead of a dozen digits, a 2D code can hold hundreds of characters, URLs, structured data, or even short plain-text messages.
2D codes generally fall into two groups. Matrix codes, such as QR, Data Matrix, Aztec and Micro QR, arrange black and white modules (the small squares you see) in a grid, with finder patterns that tell the scanner where the symbol is and how it is oriented. Stacked codes, of which PDF417 is the main example, look like several ordinary barcodes piled on top of each other, with the data split into rows.
Because the data is spread over an area rather than a line, 2D codes also include built-in error correction. If part of the code is smudged, torn or covered, the scanner can often still reconstruct the original data. The exact amount of redundancy varies by code type and by the error-correction level chosen when the code is generated.
Data Matrix: small, dense, made for tiny surfaces
How it looks
A Data Matrix code is a square (occasionally rectangular) grid of black and white modules with a distinctive border: two solid adjacent edges forming an "L" shape, and alternating dark and light cells along the other two edges. That L-shaped finder pattern is the visual signature of Data Matrix. There are no large squares in the corners like a QR code, which is why people sometimes mistake it for "a QR code with something wrong with it."
What it is good at
Data Matrix is the specialist for small marks. Its design lets it pack a useful amount of data into a physically tiny symbol, and it can still be read when printed very small, etched, engraved, or marked directly onto a part using industrial methods such as dot peening or laser marking. That is why it is widely used to mark electronic components, small mechanical parts, and individual items in manufacturing and logistics. In healthcare, small Data Matrix codes are commonly used on the packaging of medicines to carry product identifiers and batch or expiry information, a practice encouraged by traceability regulations in many regions, though the specifics vary by country.
Two properties drive this. First, Data Matrix is more space-efficient than QR for the same data at small sizes, partly because it needs less quiet zone (the blank margin around the code). Second, its finder pattern is integrated into the border rather than occupying internal space, leaving more room for data.
Where it falls short
Ordinary consumer phones do not always read Data Matrix out of the box. Many modern camera apps and third-party scanner apps handle it fine, but support is less universal than for QR. If your audience is the general public and you want a code anyone can scan without thinking, Data Matrix is rarely the first choice.
Aztec: compact and tolerant of rough surfaces
How it looks
An Aztec code is a square grid built around a bullseye: a central pattern of concentric squares that serves as the finder. The data radiates outward from that centre in layers. The largest variant fills the whole square; smaller versions use a compact bullseye. Visually it reads as a dense square with a target in the middle.
What it is good at
Aztec was designed for situations where printing conditions are poor and space is tight, and where the code will be read by dedicated scanners rather than casual phone cameras. Its most visible everyday use is transport and event tickets: boarding passes, rail tickets, and admission tickets frequently carry an Aztec code. The reason is practical. Aztec can work with a very small quiet zone, so it can be placed close to text and graphics on a crowded ticket. It also offers strong error correction, which matters for tickets that get folded, creased, crumpled in a pocket, or printed on low-quality paper at home. A damaged Aztec code can often still be read because the scanner can rebuild the data from the redundant layers.
Aztec also supports a mode that requires no external reference data, meaning the symbol is self-contained even at high error-correction levels, which suits applications where the scanner must recover the data from a badly degraded print.
Where it falls short
Like Data Matrix, Aztec is not universally supported by default phone camera apps. It is a code you typically encounter on tickets rather than one you generate for your own marketing, precisely because ticketing systems use dedicated scanners that read it reliably.
PDF417: the stacked workhorse on documents and cards
How it looks
PDF417 looks nothing like a matrix code. It is a rectangular symbol made of stacked rows of short, one-dimensional-style bar patterns, often with start and stop patterns at the ends of each row. It can be tall and narrow or wide, depending on how the data is distributed across rows. If you have a driving licence, an ID card, or a shipping label from a courier, look at the back: the long rectangular barcode there is very likely a PDF417.
What it is good at
PDF417's strength is capacity and structure. It can store well over a kilobyte of data, far more than a typical QR code of comparable size, and it can encode binary data such as compressed records. That makes it suitable for documents where a machine-readable copy of the printed information is needed: driving licences in many jurisdictions encode personal data in a PDF417 symbol, and transport and logistics documents use it for routing and manifest information. Because it is a stacked code, it also has a natural aspect ratio that fits the shape of a card or a document strip, something square matrix codes do not.
Its error correction is selectable, and at high levels it can survive significant damage, which matters for documents that are handled constantly.
Where it falls short
PDF417 is physically large for the data it carries compared with matrix codes. It is also slower for a phone camera to read, since the scanner must reconstruct rows. It is a code for documents read by purpose-built or well-configured scanners, not a code you put on a poster and hope a passer-by will snap.
Micro QR: QR's smallest sibling
Micro QR is exactly what the name suggests: a miniature version of the QR code with only one position-detection pattern instead of three. That saves a lot of space, and Micro QR symbols come in very small sizes with limited data capacity, from a handful of characters up to a couple of dozen or so depending on the variant and error-correction level. It is used where a full QR code simply will not fit, such as on tiny printed components or very small labels. The trade-off is capacity and, in some variants, weaker error correction. It also shares QR's general look, so it is recognisable as a QR-family code, but not every scanner reads every Micro QR variant, and it is far less common in the wild than standard QR.
Why QR remains the default for phone cameras
QR's dominance is not because it is technically superior in every dimension. It is not: Data Matrix beats it at small sizes, Aztec tolerates worse print quality in some respects, PDF417 holds more data. QR wins on three practical grounds.
First, ubiquity of support. QR was adopted early by phone manufacturers and app ecosystems, and today essentially every modern smartphone camera reads QR codes natively, with no app install. No other 2D code has that level of guaranteed consumer-side support. If you print a QR code, you can assume almost anyone can scan it; with the others, you cannot.
Second, finder patterns built for cameras. QR's three large corner squares let the camera locate and orientation-correct the symbol very quickly, even from an angle, at a distance, or in motion. That matters enormously when the "scanner" is a person pointing a phone in a supermarket aisle rather than an industrial reader fixed a few centimetres from the code.
Third, ecosystem and tooling. QR generation is free, standardised, and built into countless tools, including free web generators. Marketing, payments, menus, Wi-Fi sharing, event check-ins: the whole consumer workflow assumes QR, which reinforces its position.
The lesson is not that QR is best, but that QR is best for the case where an unknown person with an unknown phone scans a code you printed. Where the scanner is controlled, dedicated, or industrial, other codes often fit better.
Choosing the right code: a practical guide
A simple way to decide, in prose rather than a table:
- You want the general public to scan it with a phone. Use QR. Full stop. Menus, posters, packaging, business cards, links of any kind.
- You are marking very small items, parts, or medicine packaging, and reading happens with industrial scanners or capable apps. Use Data Matrix. It is the standard choice for small direct-part marking and small-item labelling.
- You are printing tickets that will be folded, crumpled, or printed at home, and scanned by your own ticketing equipment. Consider Aztec. Its small quiet zone and strong error correction suit rough handling.
- You need to embed a chunk of structured data on a document or card, read by document scanners. Use PDF417. It is the established format for identity documents, licences, and shipping paperwork.
- You need a QR-like code but the surface is tiny and the payload is a few characters. Look at Micro QR, after confirming your scanners support the variant you plan to use.
One more consideration: generation and verification. Whichever code you choose, test it with the actual scanners and phones your audience will use, under realistic conditions, printed at real size on real material. A code that scans perfectly on screen can fail on a curved bottle or a textured card.
Conclusion
Data Matrix, Aztec, PDF417 and Micro QR are not competitors trying to replace QR; they are specialists. Data Matrix owns the world of tiny marks and small-item labelling. Aztec thrives on tickets and rough print conditions. PDF417 carries structured data on documents and cards. Micro QR squeezes into spaces QR cannot. QR's enduring popularity rests on universal phone support, fast camera-friendly scanning, and a mature ecosystem, which is exactly what most everyday code generation needs. Match the code to the surface, the data, and above all the scanner, and the choice usually makes itself.