QR Code Link Rot: Why Printed Codes Stop Working
QR code link rot is what happens when the page or short link behind a printed code disappears while the pattern stays readable. We found no study that measured how often this happens to QR codes, but studies of ordinary web links show the pace. Pew Research Center found that 38% of webpages that existed in 2013 were no longer accessible in 2023, and Harvard researchers found that about a quarter of the deep links in New York Times articles from 1996 to mid-2019 had rotted.
Those numbers describe web links, not QR codes, and the studies date from 2014, 2021 and 2024, so read them as background, not as current rates.
Facts checked on 17 October 2026.
What four sources found
The last column says how far each result carries over to QR codes, which none of them studied.
| Source | What was checked | What it found | Carries over to a printed code? |
|---|---|---|---|
| Pew Research Center, May 2024 (data to October 2023) | Just under 1 million Common Crawl webpages, 2013 to 2023 | 38% of 2013 pages and 8% of 2023 pages inaccessible; 25% of all pages | Partly. A page counted only if it returned one of nine error codes or its domain no longer resolved, so pages that load with changed content are missed |
| Berkman Klein Center, Harvard, 2021 (full text) | 2,283,445 links to other sites in 553,693 New York Times articles; 72% were deep links | 25% of deep links dead: 6% of those from 2018, 43% from 2008, 72% from 1998. 53% of articles with deep links had a dead one | Closest on age and structure, but news links, not print |
| Harvard Law Review Forum, 2014 | URLs in US Supreme Court opinions and three law journals | 50% of URLs in opinions and over 70% in journals no longer gave the cited information | Closest to print, but legal, and the count includes pages that load but changed |
| Google Developers Blog, July 2024, updated August 2025 | Not a study: Google's notice ending its goo.gl short links | Links with no late-2024 activity set to be switched off; the rest kept | Direct, for the short-link layer of dynamic codes |
Reading the numbers for a printed code
Age drives it. In the Harvard data, 6% of deep links from 2018 had rotted against 72% from 1998; Pew's 38% for 2013 pages against 8% for 2023 pages shows the same slope. Deep links are links with a path to a specific page, like a menu page. A one-week flyer faces little of this; packaging, a plaque or a manual may be scanned years later.
Usually the page goes, not the site. Pew reported that in most cases an individual page was deleted on an otherwise functional website. A printed code can fail while the home page opens normally, so "is the site up?" does not test the code.
A page that loads can be the wrong page. The Forum authors counted dead links and links that still worked but no longer showed the cited material. Of 353 Supreme Court links returning a normal 200 response, the text says only 76% still led to it, though the paper's own table counts 277 of 353 (78.5%). The New York Times study found that 13% of 4,500 sampled "intact" links had drifted significantly.
Sources differ on the legal figure. The Forum paper notes that an earlier study of 1996 to 2010 Supreme Court opinions found 29% of links "invalid", because it counted only links that returned errors. Read 50% as dead-or-changed and 29% as dead-only; the authors also name elapsed time and two extra court terms.
The short link can disappear too
A dynamic code holds a short link on the provider's domain, and the provider's server forwards each scan. If the provider closes, the page can be healthy and the code still dead. One documented case is Google's goo.gl. Google stopped accepting new links in 2018 but kept serving existing ones. In July 2024 it said more than 99% of them had no activity in the last month and would stop responding after 25 August 2025.
In its update of 1 August 2025, Google changed course: only links with no activity in late 2024, already redirected to a deactivation notice, were set to stop. In its words, "All other goo.gl links will be preserved and will continue to function as normal." The post also acknowledged that such links are "embedded in countless documents, videos, posts and more".
So not every goo.gl link was set to stop, and the ones that were had been chosen by recent activity; ask any provider whether a code nobody scans can be retired. The Forum authors made the general point about citations in 2014: "The link, a URL, points to a resource hosted by a third party." A dynamic code points at a third-party resource by design.
A checklist for before and after printing
Before printing
- Point the code at a page you control, at an address meant to last (
/menu, not/summer-sale-v3). For products covered by EU rules, how long a QR code link must keep working lists the required periods. - Choose static or dynamic on purpose. A static code has no provider to switch it off, but a dead page means a reprint; a dynamic code can be re-pointed. Dynamic vs Static QR Codes compares them.
- Or encode a stable address on a domain you own and redirect it yourself when the page moves. No short-link service is involved, but you must keep the domain renewed.
- Keep a register: one row per code with short link, destination, print location, quantity, print date and owner.
After printing
- Every quarter, and after any site redesign, scan a printed code from a phone and confirm the page still matches what the print promises (price, date, offer).
- When a page moves, redirect the old address to the closest replacement instead of deleting it.
- To see the address an old code encodes, upload a photo to the free QR code reader. It decodes in your browser and shows the address without opening it; for a short link, that is the short link, not the page behind it.
- If a code is dead, re-point it, or sticker over it and reprint. If the cause is a lapsed account rather than a dead page, see Do QR Codes Expire?. Long-lived print such as product packaging needs this routine most.
Dynamic codes and the short-link dependency
A dynamic code encodes a short link, so you can change the destination after printing and the next scan goes to the new address. At QR Maker AI you create the code once and edit its destination later, which answers the failure Pew describes: a removed page on a working site. Only URL codes can be dynamic; Wi-Fi, vCard and the other content types are static.
It also adds a dependency. The short link sits on the provider's domain, qrmk.cc in our case, so those printed codes work only while the service runs, the position goo.gl links were in. The Business plan includes one custom short-link domain, a subdomain you own such as go.example.com (pricing). That puts your domain name on the link but keeps the provider in the path; you must keep the domain renewed, and changing it later makes codes printed with the old one stop working.
Three questions to ask a provider before you print
- What happens to my printed codes if the trial ends, a payment fails or I cancel, and can a code nobody scans be retired?
- Can I export my codes with their destinations and scan history?
- Can the short link use a domain I own?
Get the answers in writing. QR Maker AI's answers, as of October 2026:
- Lapse and removal: cancelling does not stop codes redirecting; paid features pause when the paid period ends. After a Business plan expires there is a 30-day grace period with renewal reminders, and codes on a custom domain keep redirecting during it. Deleting or disabling a code stops it.
- Export: on Business, the REST API lists codes and scans can be exported as CSV, up to 50,000 rows. The
qrmk.cclinks stay on our domain. - Own domain: Business only, one per account, on a subdomain.
- Uptime: our terms include no uptime commitment.
FAQ
What is QR code link rot?
Link rot is a URL that no longer serves any content. The Forum authors separate it from reference rot, where a link still works but the information is gone or changed. Both can hit a printed code.
What percentage of QR codes stop working?
We found no source that measured it, so any figure for QR codes would be invented. The nearest data are the web-link studies above, which measure pages and links, not codes.
Can I change where a printed QR code points without reprinting?
Only if the destination can be changed after printing: a dynamic code, or a static code that encodes an address on your own domain that you can redirect. A static code that encodes the destination itself cannot be changed.
Sources
- Pew Research Center — When Online Content Disappears. Published 17 May 2024. Accessed 17 October 2026.
- Library Innovation Lab, Harvard Law School (via Harvard DASH) — The Paper of Record Meets an Ephemeral Web, full text. 2021 (no day given on the document). Accessed 17 October 2026.
- Berkman Klein Center for Internet & Society, Harvard — The Paper of Record Meets an Ephemeral Web. Published 17 May 2021; page updated 31 July 2023. Accessed 17 October 2026.
- Harvard Law Review Forum — Perma: Scoping and Addressing the Problem of Link and Reference Rot in Legal Citations. Published 17 March 2014. Accessed 17 October 2026.
- Google Developers Blog — Google URL Shortener links will no longer be available. Published 18 July 2024; updated 1 August 2025. Accessed 17 October 2026.