| Group | Percentage |
|---|---|
| of removed charities' old web addresses now serve gambling | 5.2% |
| of the addresses that still load anything serve gambling | 14.6% |
Source: Digital Royalty analysis, data collected 18 November 2026
Download dataComment
“When a charity closes nobody really thinks about the web address, it just runs out and someone else picks it up, and every link the charity built up over the years now points at whatever they put there, which for about 1 in 20 of the ones we checked is gambling, some of it casinos that aren't even on GamStop. If you're winding up a charity keep paying for the domain for a few years, it's a tiny cost, and if your site has a links page it's worth checking where those links actually go now.”
What we found
When a charity in England and Wales closes or merges, the Charity Commission removes it from the register. Its website address carries on existing, and someone has to decide what happens to it. We wanted to know where those addresses lead now.
We took the charities that listed a website on the register in December 2022 and had been removed by 15 October 2026. That gave 5,967 distinct web addresses. We drew a random sample of 1,000 and checked them on 16 October, 6 November and again on 18 November 2026. The figures on this page are from 18 November.
52 of the 1,000 (5.2%) served gambling, about 1 in 20. The 95% confidence interval is 4.0% to 6.8%, which puts the number across the whole group at about 310 addresses, somewhere between about 240 and 400. When we fetched the 52 again about an hour later, all 52 were still serving gambling.
Most of the 1,000 addresses (645) no longer load at all, and 434 weren’t registered at our October and November registry lookups, so anyone could buy them. Of the 355 addresses that still serve something, 52 (14.6%) serve gambling, about one in seven.
What the addresses serve now
| What the address served on 18 November 2026 | Addresses | Share | 95% CI |
|---|---|---|---|
| Nothing (no address, no response, error or not found) | 645 | 64.5% | 61.5% to 67.4% |
| The charity, or a successor body | 162 | 16.2% | 14.0% to 18.6% |
| Parked, for sale or a holding page | 69 | 6.9% | 5.5% to 8.6% |
| Gambling | 52 | 5.2% | 4.0% to 6.8% |
| Blocked our check, or couldn’t be read | 34 | 3.4% | 2.4% to 4.7% |
| An unrelated live site | 26 | 2.6% | 1.8% to 3.8% |
| Redirects somewhere else | 8 | 0.8% | 0.4% to 1.6% |
| Other spam (adult or foreign-language) | 4 | 0.4% | 0.2% to 1.0% |
“The charity, or a successor body” includes charities that converted to a new legal form under a different address, closure notices, and redirects to the organisation that took the work on.
Of the 52 gambling pages, 23 are aimed at UK players by their title (they mention the UK, pounds, GamStop or bookmakers). 19 are slot or betting pages in other languages, mostly Asian, and the other 10 are a mix of general casino pages and other foreign-language ones.
The count moves
The same 1,000 addresses gave 49 gambling pages on 16 October, 45 on 6 November and 52 on 18 November. Part of that is real churn: pages go up and come down. Part is our check. Some gambling sites sit behind a bot check that sometimes lets an automated visit through and sometimes doesn’t. Of the 12 addresses that were gambling on 18 November but not on 6 November, 9 had been blocked on 6 November: 8 by a bot check, and one by a hand classification that no longer held. Every count on this page is what we could see on the day, so treat each one as a floor.
When the addresses changed hands
We looked up each gambling address in the domain registries. 41 of the 52 (79%) had been registered afresh since December 2022, when the charity still listed the address on the register. The registration dates run from October 2023 to 24 October 2026.
Most of it is recent. 29 of the 52 were registered in 2026 and 11 in 2025. Of the whole sample, 189 addresses had been registered afresh since December 2022, and 45 of those (23.8%) now serve gambling or other spam.
In 6 of the 41, the new registration came before the Commission removed the charity. For a time, the register listed the charity at an address that someone else had already registered. Across all 41, the median time from the charity’s removal to the new registration was 479 days.
The other 11 gambling addresses show registration dates from before 2023. A domain bought at auction before it expires keeps its original date, so these may also have changed hands. We can’t tell from the registry, so we don’t count them as re-registered.
Charity size
| Charity income in the 2022 register | Addresses checked | Serving gambling | Share | 95% CI |
|---|---|---|---|---|
| Under £10,000 | 495 | 21 | 4.2% | 2.8% to 6.4% |
| £10,000 to £100,000 | 295 | 19 | 6.4% | 4.2% to 9.8% |
| £100,000 to £1 million | 135 | 10 | 7.4% | 4.1% to 13.1% |
| £1 million or more | 22 | 1 | 4.5% | 0.8% to 21.8% |
| Not stated | 53 | 1 | 1.9% | 0.3% to 9.9% |
Addresses of charities with £10,000 to £1 million of income were hit more often than those of the smallest charities. The intervals overlap and the top band has 22 addresses, so read this as a direction rather than a set of rates.
We also tried grouping the 52 by the cause in the old charity’s name. Keyword groups built from names are too crude, and the numbers too small, for us to report a pattern, so we don’t.
Sites outside GamStop
GamStop is the free UK scheme that lets a person block themselves from online gambling. Every online gambling business licensed in Great Britain has to take part. A site that advertises itself as “not on GamStop” is one the scheme doesn’t cover.
At least 14 of the 52 gambling pages (26.9%, 95% CI 16.8% to 40.3%) promote casinos or bookmakers that are not on GamStop. They do it from addresses that a charity used until recently.
If gambling is causing you or someone close to you harm, the National Gambling Helpline is free and open 24 hours a day on 0808 8020 133. Unlicensed gambling sites aimed at people in Great Britain can be reported to the Gambling Commission.
Who still links to them
Old links outlast the charity. Councils, NHS services, universities and schools link to small charities from directories, support pages and resource lists, and nobody goes back to check them when a charity closes.
We took the 49 addresses that were gambling on 16 October and pulled the public-sector pages linking to them from a commercial backlink index (31 October 2026). On 18 November we loaded every one of those pages and checked that the link was still there.
14 of the 18 addresses with index results were still linked from at least one UK public-sector page: 122 pages across 57 public bodies. One of the 14, now dead, accounts for 63 of those pages on its own. The 11 that were serving gambling that same day are linked from 47 pages across 22 public bodies.
| Linking body | Bodies | Pages | Linked addresses (of the 14) |
|---|---|---|---|
| University, college or research body | 43 | 104 | 10 |
| Council or regional authority | 5 | 7 | 4 |
| NHS | 4 | 6 | 2 |
| School | 3 | 3 | 1 |
| Central government or a national body | 2 | 2 | 2 |
Two of the linked addresses were selling casinos “not on GamStop” on 18 November. One of them is linked from three NHS bodies and a council, on six pages, with the old charity’s name as the link text, so nothing on the linking page warns a reader where it now goes. The other is linked from a university page.
These link counts are a floor. No index has every page, a further 114 pages refused our check and aren’t counted, and the 16 gambling addresses that weren’t gambling on 16 October have no link data.
What this does and doesn’t show
The 52 is a floor. Some gambling pages show different content depending on who is asking. Pages that only show gambling to search engines, or only to some visitors, would look clean to our check, which came from one automated browser on one UK data-centre connection. The 34 addresses that blocked our check stay in the 1,000 and out of the 52.
The addresses belong to charities that had already closed, merged or been removed. None of this says the charities did anything wrong, and none says the bodies linking to them knew.
What charities and the people around them can do
If you’re closing or merging a charity:
- Keep the domain registered for several years after the charity closes, and point it at the successor body or at a plain page that says the charity has closed and when. Renewing a domain costs little next to what it can be used for once it lapses.
- Ask the councils, directories and partners that link to you to update or remove the link.
If your organisation links to local charities, check those links when a charity on your list closes or merges. A charity that has been removed from the register is the cue. The Commission’s register search shows removed charities and the date of removal.
Method
Method and limits
Data collected 18 November 2026
The cohort comes from two editions of the Charity Commission’s full register download: the edition of 17 December 2022 (an archived copy of the same official file, captured by the Internet Archive) and the live edition dated 15 October 2026. The live file no longer holds website addresses for removed charities, so the old address comes from the 2022 edition.
We took main charities (not linked charities) that were registered with a website in December 2022: 98,471. Of those, 8,348 had been removed by 15 October 2026. We reduced each website to its registrable domain and removed hosted platforms (social media, site builders, and gov.uk, nhs.uk, sch.uk and ac.uk addresses: 411), domains still listed by a charity that is still registered (mergers and conversions: 1,798) and values we couldn’t read (117). That left 5,967 distinct domains. We drew a seeded random sample of 1,000.
We checked each domain on 18 November 2026 between 11:05 and 11:33 UTC. The check looked the domain up in DNS, then fetched the homepage over https, then http, then the www version, with an ordinary browser user agent, a 15 second timeout and one second between domains. We classified each page from its title, its outbound links and whether it carried the charity’s name, using the same rules as our earlier study of lapsed public-body domains. Every page classed as gambling was read by hand, and 10 classifications were set by hand, each checked against the 18 November fetch and applied only while our rules still read the page the same way. A page counts only if it was fetched that day. All gambling and spam pages were fetched again between 12:18 and 12:22 UTC.
Registration dates come from the registries’ public records: Nominet’s WHOIS service for .uk addresses and RDAP for the rest, refreshed on 18 November for every gambling and spam address. For the rest of the sample the registry status is from our 16 October and 6 November lookups, because the registry limited how fast we could re-check all 1,000. Archive captures from the Internet Archive confirm, for 10 of the 11 flagged domains we could check, that the address once served the charity’s own site.
Public-sector links come from a commercial backlink index (pulled 31 October 2026) covering .gov.uk, .nhs.uk, .sch.uk, .ac.uk, .police.uk, .gov.scot, .gov.wales and .nhs.scot hosts, for the 49 addresses that were gambling on 16 October. Each linking page was fetched once on 18 November between 12:22 and 12:30 UTC and counted only if the link was in the page. The linked addresses were fetched again at 12:30 UTC.
For the comparison group we took main charities that were registered with a website in December 2022 and still registered on 15 October 2026, and applied the same domain rules. We dropped 71 addresses that were also in the removed group, which left 77,583, and drew a seeded random sample of 1,000. We checked them on 18 November 2026 between 17:24 and 17:42 UTC with the same check and the same rules, read every gambling page by hand (one classification was set by hand), and fetched each gambling page again at 20:49 and 21:01 UTC.
The intervals are Wilson 95% intervals, and the interval for the ratio between the two groups is a Koopman score interval. Estimates for the whole group are the sample share multiplied by 5,967.
Limits:
- The cohort covers removals from 18 December 2022 to 15 October 2026 only, because earlier editions of the register file weren’t checked.
- One automated fetch from one UK data-centre connection. Pages that hide gambling from some visitors would be missed, so 52 is a floor.
- Registry dates miss changes of hand that happen before a domain expires.
- “Carries the charity’s name” means the page shows the old charity’s distinctive name words. It doesn’t confirm the page is run by the charity.
- The link counts depend on what the index has crawled, so they are a floor.
- The comparison group was checked about six hours after the removed group, from the same connection, and its second fetch came about three hours after the first rather than one. Both counts are floors.
- England and Wales only. Scotland and Northern Ireland have their own regulators.
- This is a snapshot. Domains are bought and dropped every day.
- We name no charity, domain, site owner or linking body on this page. The public download is anonymised.
About this research
Digital Royalty is a UK web agency. We collected this data from public sources: the Charity Commission’s register download, the domain registries, the Internet Archive, a backlink index and the addresses themselves. We re-ran the checks on 18 November 2026 before publishing, and the figures on this page come from that run, except: the backlink index pull (31 October), the archive captures, the whole-sample registry figures (our 16 October and 6 November lookups) and the 16 October and 6 November comparison counts. The comparison with charities still registered was run the same day, after the page was first published. Nobody paid for or commissioned this study. More of our studies are on the research page.
FAQ
Does this mean these charities were hacked?
No. In most cases the charity had closed and its domain registration ran out, or it was sold. Someone else then registered the address and put new content on it. Registering an expired address is legal, and it happens to every kind of organisation’s domain.
Why do gambling sites want old charity addresses?
An address that was used for years has links pointing to it from other websites, including councils, NHS services and universities, and search engines treat those links as a sign of trust. A new owner can use that history to get gambling pages found.
How can I tell if a link I’ve got goes to a charity that has closed?
Search for the charity on the Commission’s register. If it shows as removed, open the link and check what it serves now. If it has changed, remove the link or point it at the successor body.
Can I use the data?
Yes. Download the anonymised dataset (CSV): one row per sampled address, with what it serves now, whether it promotes sites outside GamStop, the year the charity was removed, its 2022 income band, the domain’s registration status and year, and whether it was registered afresh. Domains, charity names and numbers, page titles and links are removed, and the rows are shuffled. Credit Digital Royalty with a link to this page. For the method notes, email hello@digitalroyalty.co.uk.
Is 1 in 20 a lot?
To find out, we ran the same check on charities that didn’t close. We took a second random sample of 1,000 web addresses from the same December 2022 register, this time from charities that were still registered on 15 October 2026, and checked them on 18 November 2026 in the same way.
10 of those 1,000 (1.0%, 95% CI 0.5% to 1.8%) served gambling, about 1 in 100, and all 10 still did when we fetched them again. So the old addresses of removed charities were about five times as likely to be serving gambling: 52 of 1,000 against 10 of 1,000. Allowing for chance, the gap is somewhere between about 2.7 and 10 times.
| Checked on 18 November 2026 | Removed charities | Charities still registered |
|---|---|---|
| Addresses sampled | 1,000 | 1,000 |
| Serving gambling | 52 (5.2%) | 10 (1.0%) |
| 95% CI | 4.0% to 6.8% | 0.5% to 1.8% |
| Of the addresses that still load | 52 of 355 (14.6%) | 10 of 721 (1.4%) |
Two of the 10 in the comparison group are working charity websites with betting links slipped into the page, which counts as serving gambling here too. The removed group leaves out 1,520 addresses that a charity still on the register now uses, mostly after mergers. If those were counted back in, the gap would be about four times rather than five.