Skip to main content

Info pages: publishing data about user needs and metrics

Posted by: , Posted on: - Categories: Product changes, User insights

We’ve just launched ‘info’ pages on GOV.UK. These pages aim to help anyone who publishes on GOV.UK to:

  • start content design with user needs
  • spot problems and potential improvements with existing content

An ‘info’ page contains:

  • the user needs the page is intended to meet
  • the average number of unique pageviews
  • the average number of searches started from the page
  • the top 10 search terms
  • how many comments have been sent through the ‘is there anything wrong on this page?’ link on average

The averages are calculated from data over the previous 6 weeks.

There’s an “info” page for most GOV.UK pages. To access one, just insert /info after in the URL eg the “info” page for is

More examples:

Making user needs visible

According to the first design principle, starting with user needs is important. Providing an easy way to jump from content to the underpinning needs allows content designers coming to a new topic to understand the need and build empathy with the users quicker. Publishing the GOV.UK user needs should also make the team’s work more transparent and traceable. Eventually, there shouldn’t be any GOV.UK page without a validated user need.

Working in the open

We chose to publish ‘info’ pages on GOV.UK (instead of somewhere behind a login) because (according to the tenth design priciple) it’s better to work in the open. We also wanted to make it simpler for publishers to access the data, since the data was siloed within Google Analytics, Maslow and the Feedback Explorer.

Metrics you can use

Each “info” page metric has been chosen to prompt further investigation.

Unique pageviews

This shows how popular the page is. Low pageviews could either indicate that few users have that need or that the page hasn’t been optimised correctly for search engines.

Searches started from the page

A high number of searches started from the page could indicate people aren't finding what they need. The actual search terms can hint at the source of confusion or what the missed need might be.

Number of problems reported

A high number of problem reports might indicate:

  • a technical problem (eg with a transaction)
  • general dissatisfaction with the content
  • dissatisfaction with the underlying policy

This is just the start

This is the alpha release of the “info” pages and consequently has several limitations.

Only a handful of needs have been published. While user needs have been captured during the agency and AALB transition process, the GOV.UK team has plenty of work left to do around the recording of needs and linking from the content to the needs.

There are no metrics for multi-part formats (eg guides).

It’s not possible to see how a particular page performs for a given metric in relation to other pages on the site.

Sharing and comments

Share this page


  1. Comment by Jeni Pitkin posted on

    This is brilliant, thank you!

    Our content (Department for Business, Innovation and Skills) has been on GOV.UK a while. Not all of the detailed guides are based on a user need so I'm reviewing all of them. This will help me show colleagues some of the data - very helpful.

    • Replies to Jeni Pitkin>

      Comment by Huw Pritchard posted on

      Hi Jeni - I was looking at some BIS content when I wrote my comment above, but my pint applies to most government 'marketing' How would you go about identifying the user need here, beyond 'The Minister wants it'?

  2. Comment by Dave Hallworth posted on

    This is going to be very helpful. Thank Jake

  3. Comment by Padma Gillen posted on

    Love it! Nice work guys. And great that this is publicly available information.

  4. Comment by Ross Brown posted on

    Really interesting. I've been looking through a few of the pages I've worked on in the last few months and have found some useful information, particularly searches started from one of our browse pages.

  5. Comment by Andrew Robertson posted on

    Can you confirm, is the search data complete for all pages yet? I note this is alpha and not all data is available but not clear if search is included. For example the following each states there are zero searches from the page so not clear if that is true or just not recorded yet. Thanks.

    • Replies to Andrew Robertson>

      Comment by Jake Benilov posted on

      Hi Andrew,

      Thanks for your question.

      The search average is calculated from search term that's used more than twice. The actual terms only appear on the “info” pages if they've been searched for more than 10 times in the previous 6 weeks. This filtering is done partly to reduce the risk of searches with personal information appearing, and partly to reduce the noisy long-tail that might hinder, rather than help during content design.

      For the content that you mentioned, there aren't any search terms that fit these filters.

  6. Comment by Harry Metcalfe posted on

    Very cool. Can you add a route for .json so we can get the delicious, tasty data? 🙂

    • Replies to Harry Metcalfe>

      Comment by Jake Benilov posted on

      Hi Harry,

      That's a good suggestion. We're working on making the data public - we currently don't have a robust, stable API that we can publish. Stay tuned.

      • Replies to Jake Benilov>

        Comment by David Jones posted on

        Jake didn't build the info pages following this service manual then:

        "Build an API by building with the API"

        • Replies to David Jones>

          Comment by Jake Benilov posted on

          Hi David,

          The guidance that you’ve linked to stipulates how to build APIs, not that every application has to have one. The info pages are built on top of an API but we’re not ready to make that API public yet – we’re working on it.

  7. Comment by Antony Hopker posted on

    I've been looking at website data for the best part of a year, since we started transition, and I still struggle to understand the fundamental conclusion you appear to draw from this (very impressive-looking) work: how can web traffic be any indication of its relevance?

    It could be the first thing people see in a prominent place, the only thing that makes sense to them while browsing, or something that comes up in search results even if it isn't relevant.

    I can see this is a quick win for someone wanting an instant insight to a page's performance, and it is useful if you know the size of audience you are expecting to reach, but we rely more on other data about whether we are meeting user needs (eg volume of enquiries and % of completed tasks). That usually tells us whether the web page is working, not the other way round.

    I understand you could draw some conclusions from bounce rates, exits and search refinements, but not page views, and I'd like to understand how you make use of that data.

    • Replies to Antony Hopker>

      Comment by Jake Benilov posted on

      Hi Antony,

      I agree – the presence of page views isn't necessary very significant. However, the *absence* of page views is a red flag and should prompt an investigation. To quote the GOV.UK guidance (

      "Every part of the GOV.UK website design and architecture, and every piece of published content, should meet a valid user need."

      If published content gets little or no traffic and there aren't compelling reasons for it to be on GOV.UK (eg meeting regulatory requirements), that raises questions whether the user need has been correctly identified. As GOV.UK has now grown well beyond 100K pieces of content, this question has to be asked more and more often.

  8. Comment by David Heath posted on

    Really great, and fascinating to see this data for different services.

    For the 'problem reports' metric, did you consider reporting a figure relative to the number of page views, rather than an absolute number? It might allow a more meaningful comparison of this metric between different pages, and also a comparison for a single service over time which accounts for seasonal fluctuations in service usage.

    • Replies to David Heath>

      Comment by Jake Benilov posted on

      Hi David,

      Yes, we originally intended to do this but ran out of time. We will aim to deliver this feature in the next few months.

  9. Comment by Huw Pritchard posted on

    Hi - do you apply user needs to departmental press releases and so on.

    Just tried /info on and a few other similar pages and shows no user need

  10. Comment by Paul Seiler posted on

    Appreciated your sharing and the direction you are taking this.
    FYI the following link is broken, maybe because the parent (i.e. non-info) page is down/unavailable)