Monday 4 November 2013

Universal Credit and GDS – think twice


"Agile" is not a silver bullet

Universal Credit is a damsel in distress
but the Government Digital Service is not a white knight


A Department for Work and Pensions (DWP) report has been leaked to the Guardian, please see Universal credit: £120m could be written off to rescue welfare reform.

Universal Credit (UC) is DWP's system to spring the poverty trap and make work pay.

UC is in a mess. Hundreds of millions of pounds have been spent on developing the system and there is very little if anything to show for it.

What to do?

According to the leaked report, ministers and officials (and contractors?) will make a decision in mid-November. There are, again according to the leaked report, just two choices:
AStick with the existing contractors and make UC work.
BStart again with new contractors, using a more "web-based" approach to the development of UC.
Option A is said to be "not achievable within the preferred timescales", "unrealistic" and "vulnerable to security flaws". It is also said to offer poor value for money.

Option B is said to be "unproven ... at this scale" (21 million claimants), only 100 claimants would be on the newly written system by the summer of 2014 and until then ministers would have "no idea" if it would work.

To any rational person, the conclusion must surely be neither of the above. Either option, A or B, would be indefensible, unbusinesslike and irresponsible. On the evidence available, anyone choosing either A or B would be guilty of misfeasance in public office.

The Guardian article says "ministers may order both plans to be pursued at the same time and wait to see what happens after six months". A and B? That is presumably a court jester's way of indicating the absurdity of the situation.

The article finishes by saying that DWP still insist, against the odds, that UC will "bring a £38bn benefit to society".

----------

It is not the case that there are only two options available. That is an inaccurate, false way to describe the situation. There is the option of neither A nor B.

----------

It is also not the case that ministers would have "no idea" at the outset whether option B would work. They have already been warned that it is untested at the scale of UC. There are in addition all the reasons given below for ministers to be sceptical.

----------

What on earth is a "web-based" approach?

According to the Guardian article: "Sources working on the programme say Cabinet Office minister Francis Maude, who is responsible for the government's digital team [i.e. GDS, the Government Digital Service], is in favour of the fresh web plan".

GDS are noisy advocates of the so-called "agile" systems development methodology and they orchestrate a brand-building PR campaign casting themselves as the engagingly raffish, dynamic, modern champions, succeeding where fuddy-duddy traditionalists fail. It's not clear why they expect anyone else to be enthusiastic about "agile" when you consider that:
1Despite GDS's agile approach, the G-Cloud CloudStore was "temporarily unavailable" to its users for four days last week. A four-day outage in a relatively tiny system like CloudStore is one thing. Four days without UC could be a disaster.
2GDS have started their assisted digital project several times over the past two years but despite that agility there is still no sign of progress. Just like DWP, they may try to press the reset button repeatedly with UC.
3A GDS-style web-based UC would be no more immune to security problems than the alternative.
4Working with them on an individual electoral registration (IER) data-mining pilot, the Electoral Commission found that GDS:
  • Caused delays.
  • Made it impossible to assess the results of their pilot by changing procedures in mid-stream.
  • Failed to support some participants in the pilot.
  • Failed to provide the Commission with the cost of their work on the pilot, making it impossible to say how much the pilot cost or how much live operation would cost.
  • Provided poor data specifications/inconsistent postal address specifications, leading to a failure to identify eligible voters who are not registered.
The Commission's conclusion is that a national roll-out of data-mining is not justified, it won't help IER – GDS's efforts identified both people who are already on the electoral roll and people who are ineligible to vote as needing to be prompted to register.
5GDS's identity assurance programme depends on a new pan-government "ID hub" which has been certified by no-one and which is impossibly meant to offer both anonymity and an audit trail, simultaneously.
6GDS acknowledge their responsibility for the identity assurance programme (IDAP), which was first meant to go live in the autumn of 2012 and then the spring of 2013. There was no explanation for the absence of IDAP then and there is no news of the IDAP trial which was meant to be conducted with HMRC last month, October 2013. On 21 January 2013 GDS held an event called The future is here at which they announced that they had 400 days to transform government, which might suggest that IDAP should be live in February 2014 but there are no guarantees and it may yet transpire that the future was, in the event, somewhere else all the time.
7The privacy principles which should govern GDS's digital-by-default plans for public services have still not been agreed. GDS have been quite cavalier with privacy. And their public consultation followed none of the recognised procedures.
8GDS's attempt to depict public expenditure in a series of "infographics" was described as "either an attempt to obscure the data under the guise of transparency or the work of people who have no knowledge of data visualisation ...The charts in every case are either inappropriate for the data or appropriate but ineptly designed". Among other things, the charts omitted interest on the national debt.
9GDS promise that their development work will be "open" and say that openness is "the best way to make sure that we’re accountable for the things we build. As our design principles say, if we make things open, we make things better". They promise that but do not deliver.
10Four IT professors reviewed GDS's IT strategy and in their draft findings published on 7 January 2013 they declare it to be inadequate. Among other things the professors say: "It is appealing to hope that a radical change in digital service delivery can be accomplished simply through adoption of open source technologies, introduction of agile development practices, and contractual support for encouraging more SMEs with their high-levels of energy and diversity. However, this view is much too simplistic and highly risky".
11It's not just the opinion of the Electoral Commission and of the four professors that GDS ignore. The National Audit Office have expressed doubts about digital-by-default, so have the BBC, and so have 52 members of parliament in an early day motion – they, too, have been ignored. How would GDS avoid people with no web skills becoming excluded by default? They don't say. Nor have they made any progress on assisted digital, please see 2 above. Digital-by-default is being promoted in denial of reality. Like any service organisation, GDS claim to put the users' needs first and they were even allowed to make that pitch to the Cabinet on 29 October 2013. But it looks as if UC claimants who can't use the web would be ignored by GDS.
12The repeated claim that GDS's award-winning GOV.UK has replaced all central government department websites and the websites of several agencies and arm's length bodies requires some qualification. In particular, HMRC's website has not been incorporated into GOV.UK despite claims to the contrary.
13The awards won by GOV.UK are for publishing government data. GDS's ability to cope with high-volume, complex transactions like UC which calls for quite different skills is unknown.
14The billions of pounds of savings that GDS promise depend on making a minimum of 40,000 public sector workers redundant and replacing them with computer systems. Ministers have already been warned that agile is untested at these scales. DWP may find that there is no saving to be made – they may have to pay for both the agile digital-by-default public services computer systems and the staff. The four professors warn that public services are complicated transactions. More complicated than buying a book on Amazon, for example. It may simply be impossible to replace the mature judgement of human beings with a computer system.
15The "web-based" approach includes certain fashionable components. You have to be besotted by Apple's products and you have to embrace cloud computing, please see 1 above. Cloud computing is marketed as a utility. For people struggling with fuel poverty this will hardly be a recommendation – we will not want to add IT poverty to our woes. GDS elected to host GOV.UK in the cloud. Placing the details of 21 million UC claimants in the cloud will expose their data to hackers against whom, judging by the daily stream of stories in the media, there seems to be no defence. The founder of Google has warned everyone of the dangers of cloud computing as has the Managing Director of Microsoft UK. Putting your data in the cloud means losing control of it. What do GDS have to say on this matter? Nothing.
16The "web-based" approach has its exemplars, Apple, Amazon, Google, Facebook and eBay/PayPal among them. The veneration of these companies may not be shared by the rest of the population. Their aggressive tax avoidance in some cases and their reliance on slave labour in the third world in some cases could make it politically embarrassing to pay them to act as custodians of the nation's benefits data.
17If public services are to become digital-by-default – and that is GDS's mission – then everyone must have an on-line identity. Thus the identity assurance programme, please see 5 above. To that end, GDS have appointed eight so-called "identity providers" for the UK. Everyone will be enjoined to maintain one or more personal data stores (PDSs) on the web. PDSs are being marketed by the Department for Business Innovation and Skills through their midata initiative as a way to "empower" consumers and a way to make the economy grow – they imagine that a thriving market in apps will develop advising us all what to eat and what films to watch. People should be warned that downloading an app may be little different from downloading a virus. It is not clear that consumers will be empowered by midata or that PDSs will bring economic growth with them ...
18... What they will bring, to all intents and purposes, is ID cards, without the card. It is now mandatory in the UK to register to vote, registration is to be on-line and we will identify ourselves using our PDSs. Francis Maude, the Cabinet Office Minister, talks of a new way to conduct the national census, presumably via PDSs and/or the IER electoral rolls. UC claimants in a web-based scheme would have to identify themselves on-line via their PDSs in order to claim. Does the government really want to go through the ID cards debate all over again?
The UC choice facing DWP may be presented like Beauty and the Beast.

The disgraceful performance of the big IT suppliers and the ministers and officials who are meant to be in charge of them is an obvious Beast. But there are 18 reasons at least to make DWP – or whoever is making this decision – pause before casting "agile" as Beauty.

There is a decision to make in the next few days that could waste hundreds of millions of pounds of public money while doing nothing to help the putative 21 million UC claimants. In making that decision, remember, "agile" could add to the nation's unenviable stock of expensive government IT failures just as much as the alternative.

----------

Update 5.12.13

The jester wins – in the event, they seem to have opted for A and B:

Written Ministerial Statement

Thursday 5 December 2013

THE DEPARTMENT FOR WORK AND PENSIONS

Universal Credit progress

The Secretary of State for Work and Pensions (Mr Iain Duncan Smith MP) ...

Once fully implemented, Universal Credit will account for £70 billion of benefit spending each year, and bring a £38 billion economic benefit to society over 10 years ...

- As part of the wider transformation in the development of digital services, the Department will further develop the work started by the Government Digital Services [B] to test and implement an enhanced online digital service, which will be capable of delivering the full scope of Universal Credit and make provision for all claimant types.

- Meanwhile, we will expand our current pathfinder service [A] and develop functionality so that from next summer we progressively start to take claims for Universal Credit from ...
They have chosen the unachievable, unrealistic, wasteful, unproven, no idea option.

The Statement also says: “Rightly for a programme of this scale, the Government’s priority has been, and continues to be, its safe and secure delivery. This has already been demonstrated in our approach to date”.

Null hypothesis: what that adds up to is misfeasance in public office.

Can anyone disprove that hypothesis?


Updated 15.6.18

"There is a decision to make in the next few days that could waste hundreds of millions of pounds of public money" – that's what we said getting on for five years ago in November 2013.

Now the National Audit Office (NAO) have reported.

Option A above is known at DWP as the "live service", and the NAO say that: "The Department spent £837 million on live service, making it available to single claimants nationwide and to couples and families with children in north-west England from 2015. The Department closed live service to new claims in December 2017 and expects to decommission it in July 2019".

£837 million down the misfeasance drain.

There's more in the NAO report. As ever, Tony Collins has a level-headed summary.

No comments:

Post a Comment