The British are top-drawer when it comes to fumbling high-profile IT projects. We tour the rubble as the government preps its e-government push.
-Learn about a series of spectacular public-sector IT failures - See how Prime Minister Tony Blair hopes to improve project management - Explore the British plan to use e-government to improve services Even traditional British stoicism has a breaking point and for more than half a million British citizens, that breaking point was sorely tested last summer by the introduction of a £120 million ($315 million) computer system at the country's Passport Agency. Processing times for passport applications stretched to eight weeks instead of the normal 10 days. As telephone calls went unanswered - over a million in May 1999 alone - people began turning up outside the agency's six regional offices to request their passports in person.
In Liverpool, the queues stretched for more than 100 metres. In Glasgow, over 500 people queued for 150 metres. In London, thousands queued up in the rain, forcing the agency to purchase hundreds of umbrellas to loan to angry would-be travellers. Television crews descended in droves. Britons' annual hot-spot holidays in France, Greece, Spain and Portugal are an institution. Passport renewals are a must. Adding to last summer's urgency, as it unveiled the new passport computer system, the United Kingdom also began requiring that children travel on their own passports, rather than as add-ons to their parents' papers. Consequently, the problem for many citizens was that their vacations appeared to be more imminent than the arrival of their passports. At its height, the backlog of applications reached 538,000.
For the long-suffering British public, the debacle was yet another in a long string of IT failures. Undaunted, the Blair government has embarked on the United Kingdom's highest-profile public sector IT project to date - e-government, where citizens and companies transact most of their business with central government through the Internet. The problem? IT's a huge hostage to fortune. Past government IT failures have rarely touched individuals' lives: e-government will. And, as the passport fiasco showed, if it goes wrong, then the British bulldog possesses a formidable bite.
And in the passport case, that bite reached the Parliament. Senior government ministers were forced to make lengthy explanations in the House of Commons. A change in the law was rushed through, giving two-year, free-of-charge passport extensions at any post office. As hapless ministers bungled explanatory media appearances, mass-market newspapers like The Daily Express called for Cabinet sackings - an unusually forceful terminology to use in a country where the correct protocol requires unfortunates to be called on to "consider their position". Compensation was announced for those whose travel plans were thwarted by the nonarrival of their passports.
And outside passport offices, a triage system was put in place, with agency officials scurrying along the lines dispensing advice and prioritising people into "urgent", "nonurgent" and "awfully urgent indeed, old chap" categories.
Home Secretary Jack Straw announced the recruitment of additional staff - this for a system that was supposed to reduce administrative headcount, not increase it - and promised that the situation would return to normal by September, an assurance that came as little comfort to those who had July or August vacation plans. At its nadir, Straw personally promised to move "heaven and earth" to get a passport to a woman going abroad for her honeymoon after her angry member of Parliament raised her case in the House of Commons.
The Ministry of Silly Lines
For both Siemens Business Services - the computer contractor that had developed the system - and the UK government's Home Office, the passport debacle bore an awful similarity to the shambles that had occurred a few months before. That was when Siemens' £77 million ($202 million) computerisation of the Home Office's Immigration and Nationality Directorate (IND) ran into the buffers. This system was designed to handle applications for asylum, extended stays and citizenship. Instead of processing 3000 applications a month in autumn 1998, the period prior to implementation, the number of applications processed had dropped to 995 by January 1999. The legal status of thousands was unclear. Again, queues of applicants formed - stretching for hundreds of yards on some mornings - while telephone calls went unanswered and newspaper reports spoke of sacks of unopened mail piling up in the IND's corridors in the south London suburb of Croydon.
And while the problem's initial public perception revolved around hapless refugees from conflicts overseas, the economic reality was far worse. Foreign nationals working in key jobs all over the United Kingdom were affected - the country employs a lot of software engineers from the Indian subcontinent, for example - as were the top executives of incoming foreign companies intent on establishing a British subsidiary. Media coverage quickly refocused on lost jobs and lost opportunities for investment, forcing Home Secretary Straw to add several hundred employees to the IND's staff.
For a system that was supposed to process more applications, faster and with fewer people, the awful reality was that the new system was taking longer, requiring more people to administer and processing fewer applications. Not surprisingly, the government's National Audit Office and Public Accounts Committees, the two watchdogs that monitor government spending, duly lambasted both Siemens and the civil servants in charge of the project, with Siemens being fined £4.5 million ($11.8 million) for its role in the disaster. (In fairness to Siemens, the penalty levied on it in the aftermath of the Passport Agency implementation was much smaller and reflected the role of the government's own rule-changing in exacerbating the shambles - a fine of a mere £66,000 [$173,000], despite the fact that the new system resulted in the rising cost of processing passports from £12 to £15.50 [$31 to $40].) Yet spectacular though the Passport Agency and IND failures were in terms of their public profile - there's nothing like queues of people on the television news to alert media and opposition members of Parliament (MPs) to savage the government - the systems' difficulties were relatively small fry in the fast-growing pantheon of British government computer fiascos. In fact, one of the real surprises in both imbroglios was that opposition MPs and media commentators still had the energy to lambaste the government for its information technology record - or, indeed, that they could come up with new and different synonyms for shambles, disaster, fiasco and bungle.
Curiouser and Curiouser
Take a look at some other recent British government computer fumbles. After massive computer failures in a new Social Security system installed by Andersen Consulting, for example, some 400,000 people received compensation payments of £10 ($26) each - a National Audit Office investigation had reportedly found no fewer than 1900 separate failures. The system, which was supposed to track Social Security contributions and calculate benefits owed to widows, pensioners and the unemployed, crashed within days of its debut in January 1999, resulting in 17 million contributions being unprocessed and forcing civil servants to guess at the benefits payments due to people. Even so, 160,000 pensioners were out of pocket by up to £100 ($263) per week, newspaper reports claimed.
Or consider the continuing saga of the new national air traffic control computer system being implemented by Lockheed Martin, the cost of which will have tripled in the seven years it has been in gestation, and which was deferred yet again in 1999, with the result that it may not come online until winter 2002-03 - or even later. Or the high-tech system that bar-coded patients' clinical notes in Britain's National Health Service. Implemented only in 12 hospitals (after eight years of work), the costs of the project spiralled over £32 million ($84 million) by 1998, attracting criticism from the Public Accounts Committee in a report published in January for "almost unbelievably weak management".
Or the even more expensive cancellation of a system that was to have automated benefit payments by issuing claimants magnetic swipe cards, usable in the UK's 19,000 post offices. The project, which started in 1996, was finally cancelled in late 1999 with just 205 post offices converted to the new system after an estimated expenditure of £1 billion ($2.6 billion). Far from fining the outsourcing contractor in question - ICL, a subsidiary of Fujitsu - the government promptly awarded the company a contract to automate the operations of the post offices themselves, figuring that the company must have learned something about them in the three years it spent working on the project.
But best of all, according to aficionados of the British government's forays into computing (of whom there are many: One of the country's IT magazines, Computer Weekly, even has an annual award for the biggest government computer foul-up - competition is reportedly fierce), was the 1999 House of Commons report that acted as the obituary on a top-secret defence system buried deep in a war bunker beneath London's Whitehall.
Apparently code-named Trawlerman (government officials aren't exactly forthcoming with details), the system first began to run into difficulties when officials realised that the only access to the underground bunker was through a small hatch. Everything - even mainframes - had to be built, tested, dismantled, taken through the hatch, rebuilt and tested again. And the problems didn't end there. A combination of security and compatibility problems meant that by the time the plug was pulled, says one long-standing critic of the British government's approach to IT, the computers were operating to a specification so loose that it did not include a requirement that they do the job for which they were installed.
Nor were recent breakdowns isolated incidents. "[Although] 1999 was a bumper crop of IT failures in the UK government," says the University of Manchester's Richard Heeks, author of Reinventing Government in the Information Age, "a number of those turkeys were eggs laid by the previous Conservative administration and which merely happened to hatch under the [present] Labour government."
Indeed, consider the following damning indictment from The Independent newspaper: "Problems with the Department of Social Security's multimillion-pound computer project is the latest in a string of similar disasters. Hundreds of millions of pounds have been wasted on computers in the public sector, and the mismanagement appears to continue. Two months ago, it was revealed that the department had wasted about £35 million ($92 million) on rework - dealing with mistakes. Sources within the department put the latest losses as high as £125 million ($329 million). The National Audit Office and the House of Commons' Public Accounts Committee have been frequent critics of computer management by civil servants."
The problem is those words were written in 1994, and the Department of Social Security's computer system referred to was the system that the current bug-ridden one replaced. In other words, these problems with government IT in the United Kingdom are not new, and they aren't getting any less frequent. As The Independent article went on to observe, the fact is that the UK public sector's history of flawed IT implementations stretches back many, many years.
But the problems that bog the systems down don't appear to be radically unusual or different. As Heeks says, the National Audit Office and Public Accounts Committee have for years been looking at the failures as they have occurred, and have generally come up with similar reasons for them. Heading the list: a failure to set clear objectives, a lack of senior management support and taking on projects that were technically overambitious (see "Lessons Learned?", page 76). Coupled to this list, Heeks points out, the challenges facing public sector projects are usually larger than those facing typical private sector projects; staffing and skill levels are lower because of the private sector/public sector remuneration imbalance; and government organisation structures are more conservative and less flexible than in the private sector.
Apart from better public relations management of the fiascos as they happen, there are few if any signs that the British government is learning from its experiences, Heeks adds. (Coincidentally, the interview with Heeks took place May 5, the day that the British media was revelling in still another high-profile problem: the delays in the ballot count for London's first-ever elected mayor, as hundreds of electronic vote-counting machines, being used for the first time, broke down from ingesting dust coming from the green baize tables on which the ballot papers were stacked.) But for many weary British taxpayers, the best government IT-related news in 1999 was that the civil service Central Information Technology Unit was launching a no-stone-unturned review of government IT practice, headed by an IT expert seconded from the Australian civil service, Anne Steward. In contrast to previous blame-pinning exercises, ran the hype, this analysis would actually try to formulate some best practices.
Well good news for everyone except the hapless outsourcing and contracting community, perhaps, whose trade organisation, the London-based Computing Services and Software Association, promptly announced in December 1999 the launch of its own review of what went wrong - fearing, possibly, that its members would find themselves smeared with others' failures. "We're certainly concerned about the number of projects that don't go as well as they should," concedes the association's director in charge of the review, Charles Hughes. "When things go wrong, it causes problems for the client, the general public and of course the suppliers." In particular, he notes, problems absorb association members' management time as they try to correct them, add to costs and create bad publicity.
According to Hughes, the high-powered team that the association has gathered to investigate the causes of failure includes a general who headed up the Ministry of Defence's information systems (and who now works for defence contractor British Aerospace); a former computer expert from GCHQ, the government electronic spying bureau (who now heads the IT function at Clifford Chance, one of the world's largest law firms) and a professor who has worked within the Cabinet Office. Together, they are taking testimony submissions from contractors as diverse as Andersen Consulting, EDS, IBM and Sun Microsystems.
Although its findings have yet to be released, says Hughes, the finger from his group's preliminary inquiries firmly points to basic project management principles as the main shortcoming - especially within the civil service. (Now there's a surprise.) "There's oodles of good practice about, but all too often it isn't followed - many of the problems would be avoided if the government merely followed recognised good practice," he says. What's more, he adds, there's often a focus on the computer aspect of the project to the detriment of the change management aspect of the project. "There's no such thing as an IT project in isolation," says Hughes. "It's the business activity that needs to be understood and specified - the IT project is just a part of this. And it's essential to focus on the business objectives and outcomes, not just the IT outputs."
Something Completely Different
Reassuringly, this turned out to be one of the principal conclusions of the government's own findings, contained in a report published on May 22. "A change of approach is needed," recommended the authors. "Rather than think of IT projects, the public sector needs to think in terms of projects to change the way government works, of which new IT is an important part." Officials declined to comment further, pointing merely to the views expressed in linked pronouncements by MP Minister of State, Cabinet Office Ian McCartney, in which he acknowledged that harnessing the power of IT is not always easy. "The tasks involved are very complex and fraught with risk," said McCartney. "The government has already successfully implemented a range of complex projects. However, we still need to improve performance and avoid the mistakes of the past."
Apart from the "think of the project, not the IT" recommendation, the report also suggests an organisation change: the creation of a single responsible person to oversee each project rather than the management-by-committee approach that had preceded it. Changes in the procurement process were also suggested. Startlingly, these included the importance of suppliers understanding the requirements for a new system - and only promising what they could deliver - and a firm vow by the government to manage suppliers better. "This government will not tolerate failure and repeat the mistakes of the past," said McCartney. (For its part, Hughes' industry group, the Computing Service and Software Association, says it agrees with the government's findings.) Not tolerating failure is a laudable goal but one that could prove tricky to achieve. Experts such as Marcus Pollett see the review as a welcome step in the right direction - providing that deeds follow words. And that could be difficult to determine. Pollett, a London-based writer and commentator who has followed the UK government's IT projects for some years, is scathing about both the review itself and the culture of secrecy that shrouds it and much else in the British civil service. "It's tremendously difficult to get any useful information at all from these people," he complains, referring to the Central Information Technology Unit and the Cabinet Office that oversees it. "There's a real control-freak mentality: things do leak, but it's a tremendously secretive culture."
All of which bodes ill for e-government, a bold bid to deliver all services online by 2005. "One of the key ways in which businesses have applied e-commerce techniques is in managing their relationships with customers and suppliers," says an official in the Cabinet Office. "Public sector bodies must do the same if they are to derive the same benefits in terms of reduced costs and better procurement. The effect of the public sector doing this as a whole will itself be a significant contribution to advancing e-commerce."
Not to mention cutting the cost of government. Neil Mellor, program director of the e-government division at British Telecom - to which the government is outsourcing some of the work - points to the significant savings. The savings could be as high as £13.5 billion ($35 billion), he estimates - equivalent to a cut in income tax of between three and four pence in the pound. E-procurement and electronic service delivery alone could save £6 billion ($15 billion).
And already, points out a government official, work has begun on building the Internet portal through which citizens and businesses will communicate with government. Intended to be up and running this summer, it will provide relatively static information at first, but from a single source rather than multiple Web sites. Work on the second stage of the project - building links between the portal and government IT systems, so that, say, a change of address has to be communicated only once - will begin later this year.
But, ever accident-prone, the UK government soon slipped on yet another banana skin. Announced with a huge fanfare in April, officials forgot to check the name for the portal - UK Online. It's a great name for a portal, but it also happened to be the name of a middle-ranking Internet service provider, which was less than amused to find its name hijacked. Once again, awkward questions were asked in the House - this time from the member of Parliament in whose constituency the business is based.
An accommodation was apparently reached, and UK Online continues to be the name of the e-government portal. How? On what basis? Officials remain tight-lipped. Good lord, no, they say. We can't release information like that . . .
In a report released in January, the House of Commons' Select Committee on Public Accounts examined 25 government IT projects that did not work and came up with project management lessons from each, including * Planners need to have clear project objectives and milestones.
* Projects that are part of an overall strategy must clearly fit together, and those implementing the separate projects must understand this clearly.
* Carefully examine the full implications of introducing changes to IT systems, including the impact of any policy changes (like passport laws) that may affect the demand for services.
* Contingency plans are essential to cover the risk that the system won't be delivered on time.
* Planners should consider whether programs with sweeping goals might be too ambitious to attempt in one project.
* Arrangements with subcontractors should include terms with both incentives to reach project goals and penalties if they don't.
* Delays in implementing projects place them at risk of being overtaken by technological developments.
Could It Happen Here?
The UK doesn't have a monopoly on government IT fiascos While not quite on a par with the United Kingdom's efforts, the Australian government - always loath to lag behind world's best (or worst) practice - can cite its own share of IT fiascos over recent years.
In August 1998, for instance, the Tax Office blamed intermittent problems with an interface between two systems and "sporadic and isolated" problems caused by hardware and software problems for severe delays in processing ELS (Electronic Lodgement System) returns. ATO figures showed a larger backlog than usual meant delays to well over 273,000 returns.
The same year the government denied massive staffing cuts had been to blame for problems plaguing Centrelink computer systems that allegedly caused chaos, with increases in queues and thousands of incorrect benefit payments. Some 5000 customers experienced payment problems when the major new Newstart Common Platform system was introduced in May 1998. The CPSU claimed at the time the Centrelink system had more than 800 problems and condemned the government and Centrelink for rushing into the new system without providing adequate staffing numbers and technological support.
Then in January 1999 Federal Community Services Minister Wayne Truss was forced into damage control mode after minor computer problems and glitches caused 37,000 overpayments over Christmas. Centrelink said the problem affected people who received the youth allowance or Austudy nationwide, with some people overpaid by up to $800, but was much less frank about the underlying IT problems, which it claimed were minor.
Moving on to this year, the Australian Customs Service was forced to spend three months manually processing refund-related payments under the new GST regime after experiencing significant problems in getting its systems ready. Of a couple of million entries processed by Customs each year, some 8000 were affected and required a manual workaround.
That problem came after a year's delay to another Customs system outsourced as a separate contract to EDS, the Passenger Analysis Clearance and Evaluation system (PACE), which had a contract price of $5.4 million. Changes to the design of the system and delays in getting data from the Department of Immigration and Multicultural Affairs as well as delays in the work done by EDS were blamed for holding up PACE.
But the crème de la crème of Australian government IT woes has to be the saga of the dud subs. A damming July 1999 report into the Collins-class submarines warned the combat system was so flawed it should be trashed, the vessels were noisy and vulnerable to attack and the engines constantly broke down. The report found faults with the communications system made contact between the submarines and their headquarters slow and cumbersome and the subs highly vulnerable during transmissions. Oh yes, and crucial sonar equipment - the submarines' "eyes'' - was only 25 per cent effective.
The report recommended the combat system be immediately removed from the prime contract with the submarine builder, the Australian Submarine Corporation, and then a tender be put out for a whole new off-the-shelf system. Some analysts put the cost of correcting the numerous problems detailed in the report by CSIRO chief Dr Malcolm McIntosh and former BHP managing director Mr John Prescott at $1 billion. - S. Bushell Big E-Government Plans Start SmallIn a small way, e-government is already happening. Companies can get export sales leads from embassies abroad e-mailed to their desktops, matched to their requirements. They can register new business vehicles with the UK's Driver and Vehicle Licensing Agency, and file tax returns. Various forms of advice are also available.
In the next year, government plans call for companies to be able to register for value added tax (VAT) and get truck permits issued online. After that, the pace quickens. By 2002, businesses will be able to submit their annual returns to the business registry, file for landfill tax, air passenger duty and insurance premium tax, and conduct land ownership surveys. - M. Wheatley