Skip to main content

It's all about the nodes and what lives at them

Posted by: , Posted on: - Categories: GOV.UK

To say that no-one wants to visit a government website is an over-simplification. But it's certainly true that the compulsion is significantly different from what drives people to a destination website like BBC News or Facebook, or from something customer-facing like Amazon.

"Government isn't Amazon" is hardly headline news, and may seem like stating the obvious, but treating a website as a destination in its own right, or assuming some kind of traditional customer relationship is a trap that many websites, both public and private sector, fall into again and again.

Those of us responsible for commissioning, building or designing a website are often unable to see beyond its edges (I've been just as guilty of this in the past as anyone else) - to see where it fits into people's lives and why people might be there.

People interact with government online either when they have to,  eg applying for Job Seeker's Allowance, filing a tax return, or to get information that government provides to citizens for support or statutory obligation, eg "what rights do I have?" "how do I do so-and-so?".

There are exceptions to this, such as areas of specialist content (eg professional guidance for teachers or accountants) or content that exposes the workings of government for internal and external scrutiny - both of which will be covered in future blog posts.

But for the majority of the population, the internet is the browser icon on their desktop, the starting point is Google, and they have a parking ticket to pay and are a bit grumpy about it.

People want to be in and out as quickly as they can, with as few surprises and as little faff as possible.

It's all about the nodes, aka landing pages, and what lives at them, not the hierarchy of the website that they live on.

Many commercial websites live or die by the effectiveness of their landing pages, and the same applies to government.

As has been mentioned in previous blog posts, we spent a considerable amount of project time looking at all the search data we could lay our hands on to identify what the top things people were looking for from government.

We found that 1) Google is the starting point for the significant majority of journeys, and 2) there is a relatively small number of tasks that lots of people want to do (Peter Jordan will be blogging about this in more detail soon).

People are arriving from search engines looking to complete a specific task, so we decided to build an unashamedly flat, task-focused website to help people find the 'quick do' as we termed it.

The assumption is that people are coming from a search engine, and for those who aren't we provide an autocomplete search box with a more traditional search results page as a backup - again to get people straight to the answer as quickly as possible. It's about the nodes not the network.

It's not that there isn't any Information Architecture on - there is a concept of related content to link together things like 'report a lost passport' and 'apply for a new passport'. But it's very flat, and relies on editorial choices rather than a fixed hierarchy.

The second part of this is what lives at the nodes - what are the nodes landing pages for? We realised early on that, mainly due to the constraints of one-size-fits all content management systems, many government websites tried to answer every need in the same way - with an article consisting of text and links.

We took the opposite approach, that every node (landing page, endpoint, microsite - call it what you will) should be tailored to the task at hand.

What started out as a kind of joke amongst the team ("We're building a NotCMS") became a defacto design rule.

Again, more on this in a future post...

Sharing and comments

Share this page


  1. Comment by Writing simply: language choices for the GOV.UK navigation | Government Digital Service posted on

    [...] is clearly way less important than getting the site to work properly and to the practical UX work of getting people to the content they need, but we thought you might be interested in some of the [...]

  2. Comment by Introducing the beta of GOV.UK | Government Digital Service posted on

    [...] those currently catered for by Directgov) - making them as findable, understandable and actionable as we can. We’ve built a scalable, modular open source technology platform to support them, [...]

  3. Comment by – From Alpha to Beta | Government Digital Service posted on

    [...] on a ruthlessly understanding and meeting user needs. We need to make it as easy as possible for users to complete tasks, understand information and work out what they need to do [...]

  4. Comment by Bill posted on

    Are we looking for a UK Government version of which links information from as many places as it can. Great site but it's developed over the years and take time and a lot of effort I'd imagine.

  5. Comment by Alex posted on

    Do we think there is a risk this site is designed by and for 20 - 40 year old IT literate males ?

    Or did I miss a female in the design team and comment writers - Different Anon or RT ?

  6. Comment by jeffrey petts posted on

    It’s right of course to set out the design problem in terms of what govt does and when it interacts with citizens – and also right that’s essentially a twofold need of citizens (rather than a set of wants) around required transactions with govt and informational needs about rights and obligations. Given that, the kind of design solution alpha offers makes sense.

    The caveat to that is the one that often gets discussed: broadly behavioural change content. If govt wants to get involved in that, it legitimately can and it’s no use wishing it away. I think the question is whether this kind of online activity by govt is part of the design problem for alpha – if it is, then the simplicity and clarity of the alpha solution will be quickly lost as depts start to add content to the site. But it’s also possible that it could be excluded from the alpha design problem by a strategic decision to, roughly, change behaviour by online and other means outside of alpha.

    One further point on ‘tasks and people’ (that Peter Jordan will be taking up) – that’s important too because I think alpha should commit to all govt transactions and rights etc content as a default, and use metrics on specific tasks and audiences to influence the detail of a design solution.

  7. Comment by Russell posted on


    Can you explain why's the 'Childcare' guide is better than what's already on Directgov? A search for childcare (which is nicely presented as a 'recommended link' in search results gives this page:

    ...a search on alpha gov gives tons of results - all looking the same - and includes council tax and prime minister's questions.

    • Replies to Russell>

      Comment by Matt T posted on

      One assumes they are preparing for the day that there is no direct gov... No-one is actually going to _say_ that, obviously 🙂

      • Replies to Matt T>

        Comment by Russell posted on

        ...that doesn't answer the question...

        • Replies to Russell>

          Comment by Matt T posted on

          ... did somebody say it needed to be better ? Look how shiny it is !

          In all seriousness I'd imagine the content is a mixture of finished / placeholder / experimental / indicative at the moment. Makes things hard to judge though, certainly.

  8. Comment by Richard Pope posted on

    Different Annon - you are absolutely right, gathering information is itself a task. That was the thinking behind the 'guides' concept.



    Thanks for spotting the bug with related items, I've added it to the bug list.

  9. Comment by Different Anon posted on

    "Should we make that box more obvious in the design of the page?"

    Maybe (though you might first want to sort out the bug that one of the related links on "Report a Lost Possport" is "Report a Lost Passport"...). But that's all detail. You're ignoring the big issue, which as Russell puts it above is that by focussing on "tasks" to the exclusion of all else, you're losing that wider sense in which gathering information is itself a task.

    I'm really not trying to be completely negative - there's lots that's good about Alpha. But by taking what might have been useful starting points to ridiculous extremes, by being so militant about your decision-making ("What started out as a kind of joke amongst the team... became a defacto design rule") you're throwing the context baby with the stickiness bathwater.

  10. Comment by Richard Pope posted on

    Different Anon - related things are shown in the bottom right of this page:

    Should we make that box more obvious in the design of the page?

  11. Comment by Francis Irving posted on

    Ian - my guess would be that building communities and tourism also should be done with action focussed landing pages.

    In the case of tourism, really study what tourists search for when thinking about where to go on holiday (to win customers from other countries, or parts of the country), and then what they really need when they are definitely going to travel to the local authority (or are actually there). Then create pages to server their needs.

    It is quite likely such pages would even be on another domain - nobody using them will care or notice if they are on the authority website.

  12. Comment by Ian H posted on

    Interesting concept... but isn't that only looking at half the picture?

    Take a council site for example. Much of it will inevitably be focussed on transactional services, and linking them together as you suggest undoubtedly works. But councils are also supposed to build strong communities, encourage tourism, and develop the local economy. How will that work in this IA model?

    I'm not saying this just to pick holes... I'm just really interested in this project so keen to hear how you might think that might fit in.


  13. Comment by RT posted on

    You definitely got me interested. I'm curious, what would we be able to do on that we couldn't go with a google search or a search on direct gov?

  14. Comment by Different Anon posted on

    The unspoken assumption there is that the user knows exactly what he or she wants to do before they arrive. That might be the case in many instances, but what about those who don't come with a specific task - or who do, but don't realise that there are other things that they might also need?

    As an example, if I search on DirectGov for "Register a Death", I get a link to the online tool to register a death. But I also get links to a checklist of things to think about when someone has died, information about Probate, documents I need to have to hand, arranging a funeral, bereavement allowance, and lots more. I know from experience how helpful I found it to have all that in one place.

    If I do the same search on Alpha, I get a tool to register a death and, er, that's it. I know which I'd find more useful.

  15. Comment by Adrian Short posted on

    Spot on. We need slippery websites not sticky ones.

    Get in, do the business, get out again as quickly as possible. I'm interested in exploring how many ultra-top tasks can be performed on the home page itself.

    I don't want to read your press releases.

    • Replies to Adrian Short>

      Comment by Russell posted on


      No, this is a common misconception. Sure, many people want to 'do something quickly' on (say) Directgov. But not everybody.

      There are situations where users need INFORMATION at important times in their lives: getting divorced, adopting a child, having a medical assessment to get a disability benefit, their rights when being made redundant, planning a will and so on. Nor do they want to trawl through results aimed at multi-audiences or identify themselves as (say) a member of the public and not a 'professional- post-search.

      It's easy to build a 'I lost my passport' tool but content shouldn't be neglected.

      Wikipedia wouldn't exist if people didn't invest their time in learning and understanding sometimes complex stuff.

    • Replies to Adrian Short>

      Comment by JonnyS posted on

      That assumes that the relationship between user and govt website will always be transactional.
      I think that misses an opportunity for users (and Govt)?

      Why can't we use data to predict what services will be relevant for an individual / family group and seek to solve problems before they've arisen (thus preventing demand for acute services)?

      It just requires users to trust!??!