We read the website of every charity on the three UK registers that has one on file, and publish what we find as aggregate figures with the method and the data behind them. No charity is named.
Churches and other religious bodies are excluded from the census papers; each one says how, and why. Every paper carries its denominators, its detectors' measured error rates and the aggregate table behind every figure.
48.9% of 55,300 UK charity websites run WordPress, and 13.6% of the 20,037 whose version is readable had not installed the latest security release for their branch. Platforms, versions, PHP and page builders measured in August 2026.
62.3% of UK charity websites that run cookie analytics show no consent tool; 26.4% of all readable homepages load cookie analytics with no consent banner - measured across 55,300 sites.
87.2% of UK charity domains publish no enforcing DMARC policy, 55.7% no DMARC record at all, and 28.8% no SPF record - measured across 73,687 domains.
A measured census of 73,687 charity websites across the UK. Of the 34,843 sites that engage with donations online, 59.0% show a visible way to give - and 41.0% ask but give the donor no way to act.
Between 41% and 50% of UK charity register records hold no usable website, and in a measured crawl 17.9% of recorded URLs were wrong.
A measured census of 1,233 charity websites across the Bristol area: 44.7% show a visible way to donate online, the top three platforms carry over half of offsite giving, and 28.8% of charities send donors to somebody else's website.
The Charity Commission rebuilds its register extract every night. Our pipeline checks it each morning, diffs the new edition against the last, and regenerates the register paper from the current data. Since 15 August 2026 it has recorded 62,607 changes to registered charities' watched fields.
Measured by NebulaBot's pipeline, aggregate counts only · full history in the register paper
The population is every charity on the England & Wales, Scotland and Northern Ireland registers with a usable recorded website. Each site is read once, one homepage and at most a handful of follow-on requests, by an identified crawler that honours robots.txt: NebulaBot.
Every detector is scored against a blind, hand-labelled sample and its error rate is printed with the figure it qualifies. Cross-tabulated cells below the disclosure threshold are suppressed, and the generator refuses to write a page that contains a charity's name or domain.
Every paper names the query behind each figure and ships the aggregate table it was drawn from under CC BY 4.0. Register data is used under the Open Government Licence v3.0. Quote freely, with credit.
This research is by Nebula Design, a WordPress agency near Bristol that builds websites for UK charities. We publish it because the sector deserves better data about its own websites than anyone had, and because the findings shape how we work.
If they ring true for your charity, see how we approach charity web design, browse the free Funding Finder for grants that can pay for a website, or meet the team. Questions and corrections: [email protected].