Saturday 3 September 2016

GDS, the data centre for government

"The Government Digital Service [GDS] is the digital, technology and data centre for government ... Over the next 5 years, together with departments, we will be building new digital, technology and data platforms for the whole of government ... We will make better use of data to drive continuous improvement, ensure government has the right technology and that spending is controlled".

That's what it says in the job description for a new head of service design required at GDS. You've got until 13 September 2016 to apply.

And that's not the only job available. GDS, the "data centre for government", also want data scientists to build new "data platforms for the whole of government" and to make "better use of data".

Some time in 2013, it's hard to establish when, GDS and Warwickshire County Council investigated the question of Interoperability between central and local government identity assurance schemes. The project was "conceived by Warwickshire County Council, a medium-sized local authority, and involved the participation of the Government Digital Service (GDS), a team within the UK Government’s Cabinet Office tasked with transforming government digital services, and three Identity Providers – Mydex, PayPal and Verizon".

Three years later, Mydex is no longer an "identity provider". Neither is PayPal. Nor is Verizon.

One of their findings back then, under the heading Use of a Social Media ID (p.12), was that "most users would be very reluctant to use their social media accounts with a government site, the prevailing view being that their social life is distinctly separate to doing 'business' with government. The issue of privacy and the feeling that government would be able to 'see my social life', or that government transactions would appear in their social media profiles, was of concern. That said, some users saw the benefit in forms being pre-filled with details held within their social media account".

How did GDS come to this important conclusion?

Answer, the data centre for government asked five people: "User experience testing was performed in a laboratory environment and involved 5 users on a one-to-one basis with an experienced research facilitator provided by GDS".

Now roll forward three years. GOV.UK Verify (RIP) wasn't meant to go live until it could verify-the-identity-and-therefore-register at least 90% of applicants. The registration rate remains stubbornly below that figure at about 70%. Despite that huge disparity, the data centre for government which is supposed to use data to drive continuous improvement declared GOV.UK Verify (RIP) to be live in May 2016.

In order to improve the registration rate, GDS continue to give the remaining "identity providers" access to more and more of your personal information.

GDS's 25 July 2016 blog post,Can online activity history help GOV.UK Verify [RIP] work for more people?, tells us that they have returned their attention to your social media accounts, "Facebook, PayPal, LinkedIn and others".

They undertook a new project and guess what: "Compared to the findings from 2013-2014, our recent research suggests that people appear to be becoming more amenable to using online activity verification and allowing certified companies access to their personal online accounts to acquire a verified identity that gives safer, faster access to government services".

How much more amenable? How much safer? How much faster? The data centre for government doesn't say.

How did they reach this unquantified conclusion? "The user research involved 12 one-to-one sessions with users" and "We had 86 people, from across participating organisations, testing the service using their real online accounts".

Not exactly big data, is that enough data to support GDS's otherwise unquantified conclusion? The data centre for government doesn't say.

What they do say is that "if activity from such accounts could be used for activity history, GOV.UK Verify [RIP]’s demographic coverage of the adult population overall could increase by 9%, and for the 16-25 demographic could see a potential increase of up to 38%".

Suddenly we're being fed some numbers to work with. But don't get your budding data scientist's hopes up. We have no idea how these numbers were computed. We have no idea how much confidence we can place in them ...

... except that we have been here before. On 25 January 2016, to be precise, when GDS published Estimating what proportion of the public will be able to use GOV.UK Verify [RIP]: "Looking at the overall population of UK adults, we found that at least 73% of over 16s and 78% of employed people are likely to have the evidence and footprint needed to verify their identity using GOV.UK Verify [RIP]. This will increase to at least 91% of employed people and at least 88% of over 16s by April 2016. By July, coverage will increase to 95% of employed people and 93% of over-16s".

The tool shows that by July coverage will increase to 95% of employed people

That didn't happen. There is no reason to believe that GDS's touted 38% improvement will happen either. Any data scientist who joins GDS has their work cut out.


Updated 12:44

The blog post above has been updated since publication. Some links have been added. Also, the reference to 86 people from across participating organisations being involved in testing.

Frowned upon in the world of professional journalism, this iterative primping will be instantly recognisable to GDS as "agile blogging".

Updated 9.12.16

5:12 a.m. yesterday, the Government Digital Service (GDS) tweeted the claim: "Our Standards Assurance team announce savings of £339 million through spending controls".

Presumably to support that claim, GDS provide a link to a blog post, A problem shared is money saved. There, the £339 million figure is described as "confirmed" and we learn that the confirmed savings were made in 2015-16.

The figure is unaudited and we don't have a breakdown but GDS do provide a further link, to Government saves £18.6 billion for hard working taxpayers in 2014 to 2015. That's obviously too early to provide any support for 2015-16 savings but we do learn that savings are calculated by comparing the year's figures against the same figures for 2009-10: "The government made savings of £18.6 billion in 2014 to 2015 against a 2009 to 2010 baseline".

You might think that a saving is a saving. Not for GDS. Change the baseline and you'll get a different figure.

You might think it's hard to make savings. Not for GDS. Propose a £1 billion project and let the GDS spending controls people turn you down. Savings have immediately increased by £1 billion.

12:25 a.m. yesterday, the Cabinet Office tweeted happy fifth birthday to GDS and said: "5 years of government digital transformation has saved the taxpayer over £3.56 billion". The Cabinet Office do not cite any source for this figure. And what happened to the £18.6 billion figure?

12:57 a.m. this morning, Helen Olsen Bedford tweeted: "GDS claims £4 billion savings for government IT". She cites an article on the website, GDS claims £4 billion savings for government IT.

That article provides no source but it looks as if these figures are all coming from Stephen Foreshew-Cain's How digital and technology transformation saved £1.7bn last year. Mr Foreshew-Cain replaced Mike Bracken as executive director of GDS and has now himself been replaced by Kevin Cunnington. Last time we looked at Mr Foreshew-Cain's blog post on savings we said:
… And there is no audited support for the £4 billion figure. Elsewhere, we read "in many ways GDS has been a success story ... with a claimed £1.7 billion cost savings" (Helen Margetts). After reading Government unveils £14.3 billion of savings for 2013 to 2014, turn to p.4 of End of year savings 2013 to 2014: technical note and you'll find claimed savings of £91 million + £119 million = £210 million or, to put it another way, £0.21 billion. Not £4 billion. Not even £1.7 billion. Which figure, if any, is correct? More research required, Mr Rooney. And Ms Margetts …
Savings? Think of a number …
There has been no progress in the three years since Mike Bracken told the Americans that relevant savings equivalent to 1% of the UK's GDP had been made whereas, on inspection, the correct figure looked more like 0.0138%.

GDS's claims to have made savings remain confusing and therefore unconvincing ...

... and on that basis they are not the obvious candidate to be the "data centre for government". The data centre centre for government needs to provide information that is reliable and comprehensible. GDS's figures for savings are neither.

No comments:

Post a Comment