The public web Australians
are told to rely on

Every Australian government, university and charity domain in the Tranco top million — 844 sites — checked for links that no longer go anywhere.

Scanned 2 September 2026 · 844 homepages · 12,228 links checked · census, not a sample

0.7%
of links are genuinely gone (404/410)
2.5%
were refused rather than missing — mostly 403 bot walls
8.9%
of links go through a redirect
301
of the dead links point at the site's own pages

Why this one is a census

Our general link rot report samples the top of the Australian web, because checking a page's links means sending requests to whoever that page links to, and running the full list would be over a million requests made in our name.

The public sector is small enough to measure properly. There are 1,047 .gov.au, .edu.au and .org.au domains in the Tranco top-1 million list published 1 September 2026 (list ID K9QPW), and we checked every one of them — up to 15 links per homepage, using the same engine behind the checker on our homepage.

This is the whole list, not the top of it. Every figure below covers every public-sector .au domain popular enough to appear in Tranco's top million. Of 1,047 domains, 844 served us a page with links. 170 could not be fetched at all and 33 returned a page with no links — usually a bot challenge shown to our honest User-Agent. Both are excluded rather than counted as clean, which makes the numbers below conservative.

Which part of the public sector rots fastest

Broken links as a share of links checked, by sector:

Government4.08% of 4,382 links
Education3.35% of 2,804 links
Charity & non-profit2.94% of 5,042 links
SectorSitesLinks checkedBroken Sites with a dead linkCompletely clean
Charity & non-profit 347 5,042 148 60 (17.3%) 284 (81.8%)
Government 306 4,382 179 53 (17.3%) 247 (80.7%)
Education 191 2,804 94 38 (19.9%) 151 (79.1%)

What the dead links actually say

This is the most important table on the page. Of 421 links that failed, only 82 were actually missing. Most of the rest were refusals — a server declining an automated request, which says nothing about whether the page exists. Reports that quote a single "broken links" percentage without this split routinely overstate rot by several times.

Not every failed link is a page that disappeared. A 403 is usually a server refusing an automated request rather than a genuinely missing page, and we report it separately rather than folding it into the rot figure. A 404 or 410 is the real thing — a destination that is gone.

StatusWhat it meansDead linksShare
403 Forbidden — often a bot wall, not real rot 271 64.4%
404 Not found — the page is gone 82 19.5%
429 Rate limited 32 7.6%
502 Bad gateway 15 3.6%
-2 Other error 9 2.1%
999 Other error 5 1.2%
520 Other error 2 0.5%
401 Unauthorised 2 0.5%
500 Server error 1 0.2%
503 Service unavailable 1 0.2%
400 Bad request 1 0.2%

Rot they own, and rot they inherited

A broken link pointing at the site's own pages is entirely within the operator's control — somebody moved or deleted a page and nothing was updated. A broken link pointing somewhere else is a third party's decision that the site never got told about.

Their own pages301 · 71.5%
Somebody else's120 · 28.5%

Most of the dead links here point at pages on the same site — which means most of this is fixable by the people who own it, without asking anyone else for anything.

What they link to that has gone away

Hosts most often on the end of a dead link. We record hosts, never the specific pages — the question is what the public sector points at that no longer answers, not which individual URLs on other people's sites are currently broken.

HostSites linking to it with a dead link
www.linkedin.com6
twitter.com5
app-eu.readspeaker.com3
nt.gov.au2
www.sa.gov.au2
transport.vic.gov.au2
australiainstitute.org.au2
intranet.sydney.edu.au2
www.ombudsman.gov.au2
shop-ansto.myshopify.com2
app-oc.readspeaker.com2
www.acecqa.gov.au2
bsky.app2
shop.acer.org1
anzctr.org.au1

Why it matters more here

A dead link on a shop is an annoyance. A dead link on a health department, a court, a university enrolment page or a charity's help directory sits between somebody and a thing they were told to go and do. These are also the sites least able to rely on a visitor trying again somewhere else — often there is nowhere else.

The good news in the table above is that public-sector sites are, on the whole, better maintained than the Australian web at large. The bad news is that the failures cluster: a small number of sites carry most of the dead links, and they tend to be the ones with the least resourcing behind them.

Method

Source list
Every .gov.au, .edu.au and .org.au domain in the Tranco list K9QPW of 1 September 2026 — 1,047 domains.
What we fetched
One page per domain: the homepage at https://. We did not crawl deeper. Homepage links are the ones every visitor sees, and crawling entire government sites would multiply the request load on servers funded by the public.
How many links
Up to 15 per page, capped deliberately. Pages with more links than the cap are measured on the first 15, so a site with hundreds of links is judged on a slice of them.
What counts as broken
Any 4xx or 5xx response. Timeouts are counted separately and are not included in the broken figure, because a slow server is not the same as a missing page.
Known limits
A single request at a single moment from a single country. Sites that geo-block, rate limit, or serve differently to automated clients will be misread — which is why 403 is broken out rather than counted as rot. Excluded domains (170 unfetchable, 33 linkless) are not counted as clean.
Data
The full aggregate is published as au-public-sector-report.json under a CC BY 4.0 licence. Attribution to Broken Link Finder with a link back is all we ask.

Run this on your own site. The engine behind this report is the same one on our homepage — paste a URL and it will check that page's links the way we checked these. If you run one of the sites above and want to know which links we found dead, ask us and we will tell you.

← All reports