What 6 FOI requests reveal about government campaign websites

Matt Billingsley, 15 June 2026, Web accessibility, Content governance, Content strategy

Our investigation found inconsistent evidence about who government websites are for, whether they work for disabled people, and how departments measure success.

We've written before about the UK government’s £4.1 million AI Skills Hub. To us, it looked like an expensive mistake: an inaccessible website, which linked to 14 courses that already existed.

It got us thinking: how does a project like this happen?

We submitted a Freedom of Information (FOI) request to UK Research and Innovation (UKRI), asking who worked on the project, what standards were used, and how success would be measured. Their answers raised more questions. So, we sent 5 more requests to government departments to see whether this was a one-off or a pattern.

Our FoI requests covered:

Similar governance gaps appeared across all of them.

Who are these websites for?

These websites exist to inform, educate or change what people do. They are designed to help citizens do, know or get something. But you cannot change someone's behaviour if you do not understand who they are.

Some sites had no user research before launch. The Ministry of Justice (MoJ) told us “no formal user research activity was conducted specifically in relation to modernising.justice.gov.uk prior to its launch.”

UKRI said PriceWaterhouseCoopers (who delivered the AI Skills Hub) ran workshops, focus groups and user testing, and will publish a white paper including the user needs and evidence that informed it.

HMRC said its audience for Tax Help for Side Hustles is the 5.7 million small businesses and summarised their need for “clear, reliable guidance on what they needed to do and when.” However, it did not explain how that insight had been gathered.

The Home Office refused to provide any information about their campaigns Enough and Stop! Think Fraud because it would cost too much to locate it. UKRI said it could not answer our questions about Made Smarter UK because the website is owned and managed by a social enterprise.

When organisations cannot explain who informed decisions, what they learned, or how that learning shaped the final product, it becomes difficult to judge whether a site was built around users' needs or an organisation's assumptions. This lack of evidence makes it impossible to understand whether public money has been spent well.

Disabled people missing from the evidence

We also asked how departments had met disabled people's information and access needs. None of the responses showed that disabled users had been involved in research or testing before launch.

The MoJ told us: "Disabled users were not directly involved in development, and testing with assistive technologies such as screen readers was not undertaken as part of the development process."

Instead, the department said the site had been tested against WCAG 2.2 AA standards using the WAVE accessibility evaluation tool, alongside manual checks. It also created a separate "accessible version" of the website for users who might struggle with parallax scrolling of the primary site.

HMRC also told us its website was built to meet WCAG 2.2 AA standards, but did not provide evidence that disabled users had been involved in research or testing.

Accessibility standards are a baseline, not a substitute for involving disabled people in design. Content can meet technical requirements, but testing with real people helps you understand whether people can actually find, use and benefit from it. WAVE’s own help page says, “WAVE cannot tell you if your web content is accessible. Only a human can determine true accessibility,” and “The absence of errors does not mean your page is accessible or compliant.”

The ‘communications product’ loophole

Departments treated these websites as communications products rather than services, which meant they did not need to meet GOV.UK’s Service Standard.

The MoJ said that modernising.justice.gov.uk "is a public communications product which sets out government policy, rather than a transactional public service."

HMRC told us that Tax Help for Side Hustles did not follow the Service Standard, although it did go through campaign approvals and was assessed against the Government Communication Service's OASIS framework.

These websites perform an important public service. They are often the main way the government communicates with the public. If they are important enough to justify public investment, they should be held to clear standards.

A reasonable standard would include:

who the users are
what they need
whether they can access the content
what outcome the project is trying to achieve
how success will be measured

But, departments use different standards, and answers to these questions are not recorded or reported consistently.

Are these websites successful?

These sites aim to inform, educate, or change behaviour, but how do departments know whether they worked?

HMRC provided some data. It said that Tax Help for Side Hustles received 131,000 page views, achieved a 61% engagement rate and that 68% of survey respondents said the website improved their understanding of tax on side hustle income. But these metrics raise more questions than they answer. How many people completed the survey? Who were they, were they representative? Did improved understanding lead to people taking the right action? Will more people pay the correct tax as a result? Without that context, it's difficult to know what these numbers really tell us.

Elsewhere, the MoJ provided no success measures for Modernising Justice, and UKRI referred us to the white paper that has not been published yet.

Across the responses, we found little evidence of a consistent approach to defining success, measuring outcomes, or evaluating whether objectives had been achieved. Outputs, page views and sentiment are not outcomes. The real question is whether people did something differently because of the information they received. It’s impossible to tell.

What we’re campaigning for

The AI Skills Hub was not an isolated case.

Across the projects we investigated, there was no consistent public record of who the websites were for, how users were involved, whether accessibility needs were addressed, or what outcomes were achieved.

This matters because the government continues to invest taxpayers’ money in these types of projects. A new government tender has been published for an AI Skills Boost Hub project aimed at training 10 million UK workers. We hope it's built on stronger foundations than the projects we've investigated.

We're campaigning for every UK government department to publish a standardised annual report on its digital projects, including:

what was spent
who the project was for
how users were involved
whether accessibility requirements were met
what outcomes were expected
what outcomes were achieved

With transparent standards like these, the public, Parliament, and departments themselves can better understand what government is building, who it's helping, and whether it's working.

If you agree, please sign our petition: Require government departments to publish annual reports on digital projects.