Monday, August 14, 2006

Public Sector IT

Whilst at the Cabinet Office, the folks at No 10 would always remind me that there are only two news stories about Public Sector IT: (1) It cost a fortune and didn't do anything for the money (2) It's a new Big Brother and will be watching everything you do for the rest of your life I was reminded of this on Sunday by the Telegraph's "feature" on Sagas of Late Delivery. There is no shortage of data in this area, whether it's NAO reports (which are mentioned in this article), Public Accounts Committee (PAC) reports, FOI data etc but what undermines all journalism is when the facts are wrong. One of the other things I learned at the Cabinet Office and, for that matter, at the Inland Revenue and every other government department, was that you instinctively start to distrust everything you read in the papers. It's easy to do that because every so often there's a story that you know quite a lot about because it's close to what you're working on, yet the version you read in the (insert name of any paper here) is so riddled with errors that you assume that shoddy journalism must abound. After a while, you pick out the journalists that know how to balance a story and know where to get their data from and you read only their stuff; or, you just read The Register and realise that if you're going to get a slant, it might as well be an extreme one which is at least entertaining. With Public Sector IT, there's too often no need to get the data wrong to make the story work and certainly no need to get the basics wrong, which is what ticked me off about this shoddy bit of work in the Telegraph: In February, the Home Office's new payments system - developed by Accenture - was blamed for the late filing of the department's accounts. I spent a while looking this up - and made the mistake of looking for references to Adelphi and Accenture (which brings up lots of references but not closely related). Then I remembered that the Home Office uses a partnership called Sirius, headed by Fujitsu, for delivery. And that's where I found the result on Adelphi, and the fact that it wasn't developed by Accenture. Three months later [the NAO] flagged up concerns over Aspire, HMRC's new IT project - also being implemented by Accenture. I guess "project" is a pretty loose term for an outsource of the Inland Revenue's IT that was forecast to cost between £300 million and £400 million a year for between 10 and 18 years. Moreover, I'm sure CapGemini will be a little upset to see that they've been displaced from running the entire contract. The project is also not "new", it was signed in 2004. On the up side, I'd guess that they're closer with their numbers - the outsource is probably costing more than plan, but I'd like to imagine that's at least a little related to the merger of the Inland Revenue (who originally signed the contract) with HM Customs. So who's got a thing against Accenture here? The watchdog published a report on Choose and Book, a new system for hospital appointments, which has already cost £12.4bn. Now that is a lot of money! £12 billion pounds - almost the entire amount spent by the Public Sector in a single year on IT, all on a little itty bitty programme to book appointments. There certainly appear to be problems with Choose and Book, not least the fact that it is not being used by as many GPs as expected so is not booking as many appointments as it should, but it's not likely to have cost this much. If it was more than £100 million, I'd be surprised (And I'm pretty sure that £64 million was the budget cost at the beginning). Interesting, no mention of any supplier ... There have also been enormous and high-profile problems at the Child Support Agency and with the administration of tax credits, which resulted in £2bn of benefits being paid to the wrong people. Problems there have been but I'm pretty sure that the £2billion was paid to entirely the right people, but that it was overpaid - a relatively simple consequence of the legislation's requirement to adjust the payments made annually in arrears. Naturally, people who were given money spent it. The money went into the economy and, probably, boosted the retail environment and, so far at least, has not contributed much to inflation so could even be judged a good thing (in absolutely the simplest sense - after all, if the money hadn't been given away it might have been spent on another IT project that didn't work out). And no supplier here either ... Sadly the advice offered isn't all that great either, at least in my opinion: "IT evolves quickly, and so over a three-year contract the clients' needs often change also. And this can make the contract more expensive, but if you didn't adapt it the client would end up with a system three years out of date." The problem is not IT evolving - it's that requirements change; developments that take 3 years or more with no incremental deliverables are bound to suffer from updated requirements. What's needed is not "flexible contracts" but disciplined change control, a robust and challenging approach to apparently new requirements to make sure that they do, in fact, hold business value and, finally, the timely purchase of IT equipment (although that is relatively a small part of most application projects) to make sure that the right amount is bought at the right time, so as to reduce the amount of kit sitting around rusting before go live. It then goes on to confuse things: The Government has now launched a new bidding system called eAuctions, which, it claims, will reduce the £14bn annual cost of public sector IT by 20 per cent. The system allows contractors to bid online against competitors in real time. I believe this is referring to the oddly named Zanzibar (one of the Spice Islands if memory serves) - which is a buying system for commodity purchases: laptops, pencil sharpeners, paperclips and whatnot - IT or otherwise. If £14bn is the total cost of public sector IT (5 years ago I calculated it was roughly £13bn, but that was before NHS, ID cards etc so I suspect it is more than £14bn now), then a saving of 20% would be a pretty good number. But, the bulk of Public Sector spend is not on commodity goods but on resources: consultants, development resources, outsource providers etc. The good news is that I broadly agree with the last point: "Few politicians understand that making what can appear simple changes can end up costing a fortune. And of course the private sector is always willing do that work rather than point out it's a waste of money." Not perhaps that politicians don't understand it but that the private sector doesn't do enough to educate the client about what is and isn't a good spend and how the spend might be better controlled. Once government is in up to its neck, it's pretty hard to walk away without significant adverse PR or wild exit costs. The public sector doesn't need me to defend it and Accenture certainly don't (and it would be so unlike me anyway). There is plenty of other stuff out there to peruse on public sector IT, such as this, on Richard Bacon's website from a PAC report less than a month ago: By the end of March this year, the four major contractors—BT, Accenture, Fujitsu and CSC—had received only £250 million in payments on£5 billion-worth of contracts. However, their spending in what I believe to be a doomed effort to make the fundamentally flawed system work has been massive. Accenture has wisely already set aside £260 million as provisions against losses on the contract, but the remaining three companies have remained strangely silent. By the end of March this year—two and a half years in—BT had received just £1.3 million on its contract of£996 million. My best advice indicates that BT has spent more than £200 million trying to get systems up and running. That amount does not yet figure in its annual accounts. I understand that BT intends to replace its existing software supplier with Cerner, fresh from that company’s poor performance at Nuffield. However, that means that most, or perhaps all, of BT’s work in progress on the existing contract will need to be written off. Uhoh. Trouble ahead? Now, all I need is a Big Brother story. Where did I put that piece on ID cards?

3 comments:

  1. Anonymous1:02 am

    With that analysis have you ever thought of a new career as a reporter?

    ReplyDelete
  2. Anonymous11:47 am

    The cabinet office's view that
    1. "Public sector projects cost a fortune and don't do anything for the money", is unsurprising.

    I've observed (though I could be wrong,) that when the government commissions a new system, it decides on who gets it based on...

    1. How many care in the community types will be employed,
    2. How many years they will be employed for,
    3. Will they still be employed when the next general election comes.

    Just three people wrote the whole of John Lewis Direct, for a total cost for programmers/designers of say worst case, 500 grand. The programmers made a load of money, because it worked, and noone else got anything, except the management who sold it.

    The government would rather employ 500 people on a project that failed, than employ 3 people on a success, because they've got to do something with the other 497 people.

    Ian

    ReplyDelete
  3. A career as a reporter - as long as you're not suggesting that the best i could do would be the sunday sport ...

    on government's strategy for systems, I'd broadly agree if you were talking, say, tax offices or DVLA offices or whatever. Government has naturally put big offices in areas that lacked employment. The Inland Revenue effectively created the entire employment opportunity in Telford in the 80s and have done likewise in other cities.

    The interesting thing on 3 people and success versus 500 and failure is that, for the most part, the 497 others are not government people - they're employees of outsourcers, they could be onshore or offshore (although the latter in very few cases as far as I know).

    Cynicism aside, government doesn't get into a project wanting it to fail - the cost of a failing project is 5 - 10x a successful one and the PR damage can be extreme. If ID cards and the NHS blow up before the next election, will it hurt the incumbents? I'd expect so.

    ReplyDelete