QR Code Statistics 2026: First-Party Production Data

This page prioritizes statistics with a traceable source, a stated measurement window, and a clear denominator. Every QRLynx figure links to the production report that measured it. Every third-party figure links to the publisher page checked for this review. Full definitions and collection boundaries are documented on the QRLynx methodology page.

Two units recur below. Code-level figures describe the codes in each report's stated sample. Device and geography figures describe more than 100,000 scan events recorded between April 21 and July 22, 2026. A scan event and a QR code are separate units because one code can produce many scans.

Scan behavior: what a QR code actually earns

These figures come from our Scan Benchmarks report, measured on the December 2025 to April 2026 window across a sample of more than 1,000 dynamic codes and 500+ creators over 152 days. The report publishes the per-code distribution alongside its sample and measurement window.

• The median QR code gets 2 scans; the average is 44. The 22x gap between the two shows an extremely heavy-tailed distribution: a small set of codes carries most of the volume. (Source: QRLynx, 2026)
• 34% of dynamic QR codes never receive a single scan. Most are tests, saved-but-never-printed, or abandoned. (Source: QRLynx, 2026)
• The top 10% of codes receive 24+ scans; the top 1% receive 649+; only 0.5% ever exceed 1,000. (Source: QRLynx, 2026)
• 94% of codes that get scanned receive their first scan within 7 days; the median is the same day the code was created and the average is 2.4 days. (Source: QRLynx, July 2026)
• 31% of scans are repeat scans and 69% come from unique visitors, where a repeat means the same IP scanning the same code again on the same day. (Source: QRLynx, July 2026)
• Scan activity compounds with age: codes older than 3 months average 245 scans against 11 for codes under a week old. (Source: QRLynx, 2026)

Device, platform, and geography

Measured directly from more than 100,000 QR code scans recorded between 21 April and 22 July 2026. Scan volume is heavily concentrated: before any exclusion, a single QR code accounted for 24% of every scan in the window. Two accounts whose volume dominated the sample are excluded from the shares below, so these describe the broad base of deployments rather than one or two large ones. After that exclusion no single account exceeds 15% of the sample.

• 87.0% of scans come from phones, 8.8% from desktops and laptops, and 4.2% from tablets. (Source: QRLynx, July 2026)
• Among scans from phones the split is close to even: iOS 51.9% and Android 48.1%. That is a coin flip rather than the Android lead usually assumed, and it only becomes visible once the single dominant deployment is excluded: with it included, the same sample reads 58/42 in Android's favour. (Source: QRLynx, July 2026)
• Sunday is the only genuinely quiet day, taking 10.6% of scans against 14.3% for an average day. Monday through Friday take 74.6%, but five of seven days would take 71.4% even if every day were identical, so the working-week effect is modest: a weekday runs about 17% busier than a weekend day, and Saturday outscans both Monday and Tuesday. (Source: QRLynx, July 2026)
• Scanning is global, led by the United States at 14.4%, the United Arab Emirates at 12.6%, the United Kingdom at 10.1%, and India at 7.0%. Country shares are the least stable figures on this page: before dominant accounts were excluded, one Indonesian deployment alone put Indonesia at 29.9%. This measures where codes are scanned, which is not the same as where their creators are based: our Creator Behavior Report found creators concentrated in the United States. (Source: QRLynx, July 2026)

Lifespan: how long QR codes stay in use

From our QR Code Lifespan Report, a production-data survival analysis measured in July 2026.

• 50% of scanned codes are still being scanned 7 days after creation, 30% at 30 days, and 13% at 90 days. (Source: QRLynx, July 2026)
• 52% of scanned codes complete their entire scan life within one week of the first scan; 18% are scanned exactly once. (Source: QRLynx, July 2026)
• Among codes at least 90 days old, 66.9% have been scanned at least once and 13.9% were still actively scanned in the past 30 days. (Source: QRLynx, July 2026)

Creation and design: what creators actually do

From our Creator Behavior Report (December 2025 to April 2026 window) plus one figure measured fresh for this page.

• 93.8% of codes created on a platform offering both formats are dynamic rather than static. (Source: QRLynx, 2026)
• 75.2% of dynamic QR codes are edited after initial creation. Creators actually use the editability they pay for. (Source: QRLynx, 2026)
• Only 34.7% of creators change the default colors, 17.5% add a logo, and 17.1% pick a non-classic module style. (Source: QRLynx, 2026)
• Codes with a logo are more likely to be scanned at all: 68.5% record a scan versus 60.5% for plain codes. Among codes that do get scanned the median is identical, and the higher average for logo codes comes almost entirely from the top 1%, so read this as a correlation with creator intent rather than proof that a logo earns scans. (Source: QRLynx, July 2026)
• 47.1% of codes keep the default "Untitled QR Code" name, and 65.8% of creators make exactly one code and never return. (Source: QRLynx, 2026)
• 19.1% of creators organize codes into folders; Tuesday is the busiest creation day at 19.5%, and weekends account for just 11.1% of new codes. (Source: QRLynx, 2026)
• Nearly all codes on our platform ship at Error Correction Level H, but that reflects our platform default, not a creator choice: we auto-assign the level from logo size rather than asking creators to pick, so the exact split describes our configuration and not a QR-industry norm. (Source: QRLynx, July 2026)

Security and destination-review statistics

The QRLynx QR Code Security Report uses a fixed cohort of 754 URL-type QR submissions created from December 2025 through April 2026. The figures describe first-assessment screening outcomes, not the prevalence of malicious QR codes across the internet.

• 658 of 754 submissions passed initial automated screening and 96 required additional review. That is an 87.3% initial pass rate and a 12.7% review rate. A review flag is a precautionary platform result, not a confirmed malware or phishing verdict. (Source: QRLynx, 2026)
• 71.9% of submissions carried zero recorded risk indicators in the frozen report cohort. (Source: QRLynx, 2026)
• Within the historical new-domain subset, 69.2% used domains less than eight days old and 25.0% used domains registered on the submission day. Domain age is one review signal and requires context. (Source: QRLynx, 2026)
• The monthly first-assessment review rate reached 27.5% in March 2026 within the stated cohort. Monthly rates should be read with their underlying counts and fixed window. (Source: QRLynx, 2026)
• The FTC explains that QR scams can place a harmful destination behind a familiar-looking symbol and recommends checking the URL before opening it. (Source: FTC consumer alert, December 2023)
• The FBI documents fraud campaigns that use unsolicited packages and QR codes to lead recipients into credential or payment flows. (Source: FBI cyber alert, 2025)

Campaign and tracking statistics

From our study of 1,000+ QR campaigns over 12 months.

• The top 1% of QR codes drive roughly 70% of scan volume. Budget follows placement, not code count. (Source: QRLynx, 2026)
• About 63% of deployed campaign codes activate (record at least one scan), and 31% of scan events are repeat scans, so a raw scan count overstates unique reach by roughly 1.4x. (Source: QRLynx, July 2026)
• Placement is the biggest lever. Seated placements with dwell time, such as restaurant table cards and hotel in-room cards, rank highest across published placement benchmarks; we do not quote a figure because we could not trace those benchmarks to a named study. (Source: QRLynx, 2026)
• Static QR codes contributed 0% of QRLynx redirect events, by design: a static code encodes its destination directly and does not use the QRLynx redirect. A tagged destination URL and destination analytics can still measure visits, while QRLynx scan analytics apply to managed dynamic redirects. (Source: QRLynx, 2026)

The technical constants

The reference numbers behind every QR code, from the format's inventor and the standard body. These do not change with market cycles, which makes them the most durable rows on this page.

• The QR code was invented in 1994 at DENSO WAVE by Masahiro Hara, working with a single teammate for about a year and a half. (Source: DENSO WAVE)
• One symbol holds up to 7,089 numeric characters, 4,296 alphanumeric, 2,953 binary bytes, or 1,817 Kanji characters (Version 40, error correction L). A conventional barcode holds about 20 digits. (Source: DENSO WAVE)
• Error correction restores approximately 7% (L), 15% (M), 25% (Q), or 30% (H) of damaged data; DENSO WAVE notes Level M is the most frequently selected across QR use generally; our own production data above skews higher, to H, because logo-carrying codes get the highest level assigned automatically. (Source: DENSO WAVE)
• The current standard is ISO/IEC 18004:2024 (Edition 4, published August 2024, 111 pages), which replaced the now-withdrawn 2015 edition. (Source: ISO)

How a statistic qualifies for this page

A figure is included when readers can identify the publisher, date, unit, and measurement boundary. First-party QRLynx figures also link to the owning report and methodology. Market-size and adoption forecasts remain labeled as estimates or forecasts because they describe a model, not an observed production count.

Several frequently repeated figures are kept outside the main tables until a traceable origin and stable definition are available. Examples include the 323% QR growth claim, undated consumer-survey rows, and the claim that 88% of paper business cards are discarded within a week. The business card statistics investigation documents the source history of the last example. This policy also applies to QRLynx data: when two reports measure the same metric in different windows, both retain their own date and denominator.

Data for media, researchers, and AI citation

These lines are free to cite with attribution to QRLynx and the year shown. Third-party figures should be attributed to their own publishers as listed.

• "The median QR code receives 2 scans; the average is 44, a 22x gap that shows an extremely heavy-tailed scan distribution." (Source: QRLynx, 2026)
• "34% of dynamic QR codes never receive a single scan." (Source: QRLynx, 2026)
• "50% of scanned QR codes remain in use 7 days after creation, 30% at 30 days, and 13% at 90 days." (Source: QRLynx, July 2026)
• "93.8% of codes created on a platform offering both formats are dynamic; 75.2% of dynamic codes are edited after creation." (Source: QRLynx, 2026)
• "In a fixed cohort of 754 URL-type QR submissions, 658 passed initial automated screening and 96 required additional review; a review flag is not a confirmed-abuse verdict." (Source: QRLynx, 2026)
• "Mordor Intelligence estimates the global QR codes market at $15.23 billion in 2026 and projects $33.14 billion by 2031." (Source: Mordor Intelligence, 2026)
• "Global QR payment value is forecast to exceed $8 trillion by 2029." (Source: Juniper Research, 2025)

Sources

1. QRLynx Scan Benchmarks 2026 (window: Dec 2025 to Apr 2026)
2. QRLynx QR Code Lifespan Report (window: Oct 2025 to Jul 2026)
3. QRLynx Creator Behavior Report 2026
4. QRLynx QR Code Security Report 2026
5. QRLynx: How to Track Every QR Scan, 1,000+ campaigns
6. DENSO WAVE: History of the QR code
7. DENSO WAVE: Information capacity and versions
8. DENSO WAVE: Error correction feature
9. ISO/IEC 18004:2024 (Aug 2024)
10. GS1: Global industry endorsement, QR codes with GS1 standards (Jun 2024)
11. Mordor Intelligence: QR codes market (updated Feb 2026)
12. Juniper Research: QR payments to exceed $8tn by 2029 (Feb 2025)
13. eMarketer: US QR scanner forecast (Mar 2023)
14. FTC consumer alert on QR scams (Dec 2023)
15. FBI cyber alert on QR-code fraud in unsolicited packages (2025)
16. QRLynx methodology