One of the most common images in the world — a black-and-white blot scanned billions of times every day — was born of irritation. Specific, everyday, engineering irritation: a man had grown tired of running a scanner over a striped label.
At an automotive plant in the early nineties, a part passed through dozens of checkpoints. A brake disc, an engine block, a wiring harness — every element carried a barcode, and every barcode stored roughly twenty characters. Twenty characters meant a catalog number and perhaps a production date. No batch serial number, no supplier information, no tolerance specifications. To find out anything else, a worker had to consult paper documentation or a terminal connected to a database. A linear barcode was essentially a pointer: it said “look somewhere else” rather than providing the answer itself.
The problem was compounded by a Japanese peculiarity. Kanji — characters that carry whole words within them — would not fit into a one-dimensional strip. Encoding a single kanji character required more bits than a Latin letter, and the barcode became a long ribbon that the scanner read slowly and with errors. On a Toyota assembly line, where every second of downtime cost money, this became a systemic pain point. Denso — Toyota’s largest supplier of automotive components — felt that pain more sharply than anyone, because its parts made up a significant share of the flow.
In 1994, a small team working on automatic identification operated within Denso. It was led by engineer Masahiro Hara. The team was tiny — just two people. The task sounded mundane: come up with a code that held more data and could be read faster. No one was planning to reshape the world’s infrastructure — they simply needed to speed up the tracking of parts on the assembly line.
Hara and his colleague began with a simple observation: if a barcode is one-dimensional — if its stripes run in one direction — then information can be written in two dimensions, like text on a page. A matrix of black and white squares, in the same area, can hold dozens of times more data. Two-dimensional codes already existed at the time — Data Matrix, for instance — but they shared a weakness: the scanner needed time to determine where the code began and how it was oriented.
Hara’s solution — three square markers in the corners — looks so obvious today that it feels as though it must have always existed. But that simplicity rested on precise calculation. The positioning markers have a fixed ratio of black and white stripe widths: 1:1:3:1:1. That ratio is unique — it cannot be mistaken for another graphic element on packaging or a component housing. Passing across the image in any direction, the scanner instantly recognizes the pattern and determines the code’s position, scale, and angle of rotation. There are exactly three markers. Three points are enough to establish position, scale, and rotation, while the empty fourth corner serves as a clue in itself: it is always at the bottom right, letting the decoder know which way is down. Small alignment patterns handle perspective distortion — far more cheaply than a fourth marker would.
What gets me about this story is something else. NHK retellings mention go: even if a stone is placed slightly off an intersection, a player still finds it at a glance — and the scanner needed that same margin for imprecision. Hara himself recalled a different moment in an interview with Denso: looking out a train window, he saw a high-rise building whose upper floors were glazed differently from the lower ones, and realized that the code needed a single recognizable mark by which it could be identified instantly.
The second key innovation was error correction using the Reed–Solomon algorithm. The same mathematical apparatus used in compact discs to restore music from a scratched disc allowed a QR code to remain readable even if up to 30% of its surface was damaged. At the factory, that meant an oil stain, a scratch, or a partially torn-off label would not stop the assembly line.
The code was named QR — Quick Response. The name precisely described its main advantage over a linear barcode: reading speed. Denso patented the technology, and this is where the part of the story that is often retold inaccurately begins.
The popular version goes like this: the company “gave the QR code to the world,” making it public domain. Reality is more complicated and, in my view, more interesting. Denso Wave — the division that later became a separate company and took charge of the automatic-identification business — retained both the patent rights and the “QR Code” trademark. It is inaccurate to say that Denso Wave existed as an independent organization in 1994: at the time of the invention, it was part of Denso. But it was Denso Wave that later formulated the public position that changed the technology’s trajectory.
The position was this: the company would not exercise its patent rights against the use of QR codes compliant with the standard. Any scanner manufacturer, software developer, or printer could use the format free of charge. The patent remained in force — this was not a surrender of ownership, but a refusal to charge fees. The trademark was retained as well: Denso Wave controlled the use of the phrase “QR Code” to ensure it was not applied to incompatible or low-quality systems.
The difference between “giving it away” and “not charging a fee” may sound like a legal technicality, but in practice it is enormous. Denso Wave retained the ability to change its policy at any moment — and thus preserved a lever of influence. Yet over thirty years, that lever has never been used, and it was precisely this dependable consistency that allowed the market to build critical infrastructure on QR codes. No one feared that a licensing bill would arrive tomorrow.
Standardization cemented that status. In 2000, the QR code was approved as ISO/IEC 18004. An international standard is, roughly speaking, an invitation to the entire industry: the format is stable, documented, compatible. Japanese mobile operators — NTT DoCoMo, KDDI, SoftBank — were among the first to build QR-code recognition into phone firmware. By 2005, DoCoMo alone had nearly twenty million handsets with the feature, and by the middle of the 2000s, a square on an advertising poster in Japan reliably took you to a web page.
The technology was waiting for a device. The QR code had been developed for industrial laser scanners, but its real life began when a camera found its way into the mobile phone. The first camera phones appeared in Japan around the turn of the millennium — the Sharp J-SH04 in 2000 is considered one of the pioneers — and within a few years, built-in QR-code reader apps had become standard in the Japanese market.
The rest of the world lagged behind. In Europe and the United States, QR codes remained exotic for a long time — marketers slapped them on billboards, but users did not understand why they should download a separate scanner app. The turning point came when Apple, in 2017, built QR-code recognition directly into the standard iOS 11 camera app. Google did the same for Android. The one-step barrier — “download an app” — disappeared, and scanning became a default camera feature.
But between Japan’s boom in the 2000s and global recognition at the end of the 2010s, something far larger happened — and it happened in China.
In the 2010s, China made what economists call leapfrogging — “leaping over a generation.” A country with a vast population, where credit cards never became a universally used payment instrument, moved directly from cash to mobile payments. The QR code stood at the center of that transition, but reducing the story to an image on a sticker is to miss the entire structure.
Alipay (owned by Ant Group, affiliated with Alibaba) and WeChat Pay (a Tencent division) built payment ecosystems in which the QR code served as the entry point; the actual mechanics lay elsewhere. Behind the square at a street noodle vendor’s counter stood a chain: user identification through a linked bank account, transaction authorization on the platform’s servers, instant crediting of funds, notification to both sides. The QR code was the door handle, but behind the door stood the entire architecture of the super-app.
Why the QR code rather than, say, NFC — the contactless technology gaining ground in Europe at the same time? The answer is economic. NFC required a special chip in the phone and a POS terminal on the seller’s side. The terminal cost money, needed a connection to the banking network, needed maintenance. For a small trader at a market in Chengdu or Kunming, this was an insurmountable barrier. A QR code, by contrast, could be printed on a sheet of paper and taped to a wall. The seller’s cost of entry approached zero: all they needed was an Alipay or WeChat Pay account and a smartphone — not even necessarily their own, since the customer scanned the code with theirs.
By the end of the decade, Alipay and WeChat Pay together controlled more than 90% of China’s mobile-payment market. Hundreds of millions of transactions a day passed through QR codes — from taxi fares to donations for street musicians. Beggars in some cities, according to news reports, wore printed QR codes around their necks to receive alms — a detail that says more about the depth of the technology’s penetration than any statistic.
But crediting the QR code with “creating” Chinese mobile payments is a mistake of scale. The code was a necessary but insufficient condition. Without payment platforms with hundreds of millions of users, without cheap Android smartphones, without mobile internet, without the network effect in which every new WeChat user made the platform more valuable to everyone else — without all that, the QR code would have remained what it was in Japan: a convenient link to a web page. China’s leap was the result of technological readiness, market structure, and platform strategy converging.
And what if Denso Wave had decided differently? What if every scanned QR code carried a licensing fee — even a symbolic one, fractions of a cent?
That question cannot be answered with precision, but the outline of an answer is visible. Licensing fees create friction. Friction breeds alternatives. If the QR code had been burdened with payments, major players — Tencent, Alibaba, Google — would almost certainly have developed or promoted their own two-dimensional code formats. Technically, there was nothing extraordinary about it: Data Matrix, Aztec Code, and PDF417 existed alongside it. The problem lies elsewhere — in fragmentation.
Imagine a world in which Alipay uses one code format, WeChat Pay another, European retailers a third. Every seller puts three stickers on the wall. Every scanner must support three standards. Every subway ticket contains a code readable only by a particular app. It would not be a catastrophe: the market would simply slow down. Implementation costs would rise, the pace of adoption would fall, and small players would be cut off from the ecosystem before they had time to enter it.
There are precedents in the history of technology. The GIF format was encumbered by Unisys’s patent on the LZW compression algorithm, and this led to the emergence of PNG — a free alternative that ultimately displaced the original. Video codecs still live in a world of licensing wars, where H.264, H.265, VP9, and AV1 coexist, creating headaches for developers and users. The QR code escaped that fate. A single format readable by any device is an infrastructure-level advantage.
For a street vendor in Shenzhen, the difference between “a free standard” and “a licensed format” is the difference between “print a code on a printer” and “pay to connect to a compatible system.” The first option scales to hundreds of millions of points in a couple of years. The second — over a decade, if you are lucky.
COVID-19 became an unforeseen accelerator. In 2020, QR codes appeared everywhere contactless interaction was required: digital restaurant menus, electronic passes, vaccination certificates, visitor registration. The European Union used QR codes in its digital COVID certificates. Restaurants around the world replaced paper menus with stickers on tables.
The pandemic did for the global spread of the QR code what Alipay did for China: it removed the psychological barrier. Before 2020, the average European or American encountered a QR code only occasionally and often did not know how to scan it. After the pandemic, the gesture — pointing a camera at a square — became as familiar as unlocking a phone with a fingerprint.
The volume of QR-code generation multiplied many times over. Dynamic QR codes appeared, their contents changeable without reprinting — a restaurant updates its menu on a server, while the sticker on the table remains the same. Designer QR codes appeared, with logos and brand colors inside them — Reed–Solomon error correction made it possible to “damage” up to a third of the surface, and designers used that margin for aesthetics. The reserve for damage built in for dirty, scratched components went to logos and brand colors.
Ultimately, the QR code is a bridge. A physical object — a sticker, a screen, a poster, a business card, a gravestone — yes, in Japan QR codes are placed on headstones, leading to memorial pages — connects to a digital resource through a smartphone camera. The bridge is free, standardized, resilient to damage, and requires no electricity on the carrier’s side.
There is no need to romanticize Denso Wave’s decision. The company was not an altruist handing out intellectual property out of love for humanity. It was a supplier of industrial equipment, for which the format’s mass adoption meant increased sales of scanners, label printers, and software. An open standard is a strategy in which you make money not from the gate, but from what lies beyond it. Denso Wave continues to manufacture industrial readers and automatic-identification systems, and its position in that market depends directly on how widely “its” format is used.
But strategic calculation does not negate the result. The black-and-white square drawn by two engineers at an automotive-components plant thirty years ago is now printed on receipts in Nairobi, Moscow subway tickets, wine labels in Bordeaux, and the QR codes on electronic prescriptions in German pharmacies. Masahiro Hara, now nearing seventy, said in an interview that he could never have imagined anything like it — he simply wanted the scanner to read the code faster.
The manufacturing task of 1994 was solved long ago. But the square is still expanding — and the boundary between “physical” and “digital” wherever it is stuck grows thinner and thinner, until one day, perhaps, it disappears altogether.