ONS Digital

This is an internal talk I have given at the Newport and Titchfield ONS offices to introduce the Alpha project – shared here for people who missed the talk and as a future reference point.

 

For the record I’m not here to talk about a ‘next generation’ anything.

I’m also not here to dwell on the problems and limitations we are facing with the current website but I will take a minute or two to spell where things stand at the moment.

The reality is that a lot of people have worked very hard over the last three years to keep the ONS website going and it is because of these efforts that we have managed to keep to our release schedule and make a number of iterative improvements that have improved the experience of using the website for our users despite all the difficulties.

Investigations from external experts following the serious problems back in January identified a number of issues with the stability of the web platform and they advised us that development activity on the site came at significant risk. We have discussed the issues extensively and decided that only the most business critical updates will be made to the site – particularly relating to security – and that a project to build a replacement web platform would be accelerated.

The current ONS website is about to reach its third birthday – it launched in August 2011 – but it was actually conceived at least two years prior to that. Five years is ancient in web years and any website would require extensive redevelopment or replacement at that point as technology has moved on so far and user expectations have moved even further.

In the past user expectations of Government websites were pretty low but launch of GOV.UK and the formation of the Government Digital Service changed all of that in the UK. They proved that the Civil Service could deliver web services on a par with the major online companies.

The current website has never really delivered on user expectations and we have had some very vocal critics over the last two years. There are not many websites that are the subject of entire columns in the Financial Times and be the focus of a Public Administration Select Committee hearing. As recently as the start of this month both City AM and the Telegraph both ran stories explicitly criticising the website.

In an ‘every cloud…’ kind of way the recent travails have presented us with an opportunity to provide a website that meets the needs of these critics and removes the distraction of the website from what should be the real focus – our STATISTICS.

Inspired by the work of the Government Digital Service, and the lessons we’ve all learned over the last couple of years, we are trying a different approach with this project.

As recommended by GDS we are starting with an Alpha.

They define this as;

“A short phase in which you prototype solutions for your users needs. You’ll be testing with a small group of users or stakeholders, and getting early feedback about the design of the service.”

We have been collecting user insight in one form or another throughout the life of the current website (and before) and more recently have been undertaking this in a more structured manner. Through various feedback channels, user testing, website analytics and conversations with stakeholders we have a strong idea of what our users need from us and now the challenge is translating that intelligence in to a website that works for them.

Because to be clear the goal of any publicly facing website should be to meet the needs of its USERS first.

Our plan is to build a prototype based on what we know so far and get it out for user testing and consultation before the end of the year. Even before then we will be working in the open and testing elements of the ‘alpha’ as we go along and undertaking ‘show and tell’ sessions with internal and external stakeholders to ensure we remain laser focused on user needs.

We aren’t doing this alone. We have commissioned external specialists to assist us at every step of the way. The modern website is a complicated beast – especially one of the scale and complexity of the ONS site and nobody can be an expert all all of the elements needed to make it a success. We acknowledge this and are making sure we plug those gaps with the best people available to us – supplementing the know how we already have internally.

There are going to be changes in the way we do things though. I’m not going to try and explain the religious war that is agile vs waterfall but suffice to say we’ll be undertaking this project using agile delivery methods and that necessitates a new way of working.

The technical team responsible for the Alpha website will initially be provided by an external supplier and will be located in Newport to allow for a co-located, multidisciplinary team as per current web development best practice. Over time this external team will be supplemented and then replaced by an internal team also based in Newport and this team will remain in place to provide support and development to the new website.

The hard work the web support team have been doing over the last three years has been invaluable in keeping the website going and they have provided solutions to difficult problems in even more difficult circumstances but the reality is the current model is simply not sustainable in an agile environment where the team with responsibility for the website (Digital Publishing) are two hours away from the team implementing the development of it. There is too much opportunity for crossed wires, miscommunication and clashing priorities when the entire team does not share the same space and leadership.

So what is it this team is going to deliver by the end of the year?

There are two main parallel streams of work being undertaken. The first and most visible is going to be what us web geeks talk about as the User Experience piece.

This is basically the site the external user sees. Made up of the visual design but also navigation, how individual interactions with the site works

How do you download a file?

What happens when there is an error?

How does the search engine work?

What does the navigation look like as you get deeper into the site?

What does the site look like on different size screens?

..and a huge amount of other little questions that go in to making a site work..)

That last question is an important one. Already visits from users on mobile devices (mainly iPhones and iPads) is creeping up towards 20% of all our traffic to the site. This is despite the site not exactly being optimised for anything other than desktop screens.

It raises a whole host of questions we need to be able to answer – or at least come up with something we can test with users.

How do charts cope with a change in screen side and still stay valid?

What about tables? How will they cope on a portrait screen if there are multiple columns?

What about all of our downloads – can we expect people on a phone or tablet to download an Excel file or do they want a more web native offering?

This all needs to be done in a manner that is ‘simpler, clearer, faster’ (copyright Government Digital Service) but equally important it needs to be ‘accessible’ as well.

We will be working closely with the experts in the Data Visualisation centre to ensure that charts etc are rendered in the most sound way possible – or not at all. We have also invited a group of statistical inclined ‘critical friends’  to provide feedback on whether the presentation layer does justice to the data but this will not be a case of designing by committee – that way only bad things lay.

Some of the work this will showcase will look familiar to some of you – we will be implementing the taxonomy that came so close to going live in January as well as some functionality that has been discussed since the specification work of the current site. Anyone who has seen the post-Publications Hub ‘release calendar’ will also recognise our version. We are also hoping to make use of the considerable good work that went in to an improved site search for the current site that we have not been able to implement.

The goal is to present all this as a public prototype site (clearly flagged as NOT a live, production site on EVERY page!) that contains enough content to give a real feel for whether the design, navigation and search works as users expect.

Behind the scenes there will be a basic web publishing tool at this stage. Currently ONS uses Tridion for the majority of its web publishing – for those that don’t know this is a ‘content management system’ or ‘CMS’ of which there are hundreds available all of which do pretty much the same basic tasks. The trick is to find, or build, one that is flexible enough to evolve with user and business needs but also stable enough to provide the kind of reassurances (especially security) a site of the importance requires.

There is no real agreement on the best course of action here – organisations like GDS, the Guardian, OECD and the BBC built their own publishing tools specific to their needs. Other organisations like the National Archives, most of the US Federal Government and the Open University selected available ‘open source’ solutions. Our friends down at New Zealand Stats use a leading commercial CMS.

During the Alpha we will investigate our options here – while we have some specific requirements we are keen to not ‘reinvent the wheel’ and select something that is sustainable. This investigation is an element of the second, less visible stream of work we’ll be undertaking.

Beyond looking at basic web publishing tools we also need to have a hard look at whether there is an opportunity to build upon other existing ONS web platforms like the Web Data Access project or Neighbourhood Statistics. These platforms exist outside the current problems faced by the existing website and have been developed with specific user needs in mind but a proper analysis is needed to see if there is a possibility that they could be the (or part of the) engine that drives the new ‘user experience’. There are obvious benefits to the organisation in consolidating our web ‘estate’ anyway but there are also difficulties so this is a vital piece of work.

There is also the question of how we make the most of the great work of the Open Geography portal and the popular Nomis website. It isn’t a question of integrating or replacing these websites or NeSS, they are successful and well used already, but rather presenting a more coherent journey to our users and being in a position to better flag some of our wonderful resources to users who might not be quite as familiar with them as we might think or like.

There is no doubt that from a purely user focused point of view ONS joining the rest of the GSS on the GOV.UK website would offer considerable benefits despite the very real limitations of that platform for presenting statistics at the moment.

At the moment we have an ‘exemption’ from being moved into GOV.UK based on our work with the devolved administrations as well as the requirement to demonstrate our independence from Government.

The other issue, of course, is the cost and I’m not going to lie – this is going to cost.

The reality is that the website is our prime dissemination tool and we need to get it right. While we have ambitions for third parties to use and distribute our statistics to wider audiences even these organisations expect and require a usable, accessible repository of our data to underpin their own activities.

There is a saying in web circles ‘Google is our homepage’. Broadly speaking this refers to the fact that less and less people start on your frontpage – they start wherever Google sends them. The thing is Google needs to have somewhere to send them – as an organisation it is not primarily in the business of creating content of its own – it needs sites to index – it is the ultimate middleman. We need to make sure we have a site that plays nice with Google but also one that is independent from it.

Before I finish and invite questions I’d just like to promote the Digital Publishing blog. http://digitalpublishing.ons.gov.uk – this will be where we will share progress, sneak peaks of designs and make requests for assistance.