How IT Fixed London's Traffic Woes
- 07 August, 2003 11:01
In a nation known for spectacular IT failures, a new traffic reduction scheme has gained global attention for its smashing success. The secret? Strong project management applied to well-designed systems.
- Why the London traffic scheme worked when other UK government projects failed
- What city officials did to manage the project
- How a first-of-its-kind traffic relief project uses IT
In 1903, 10 years before Henry Ford began mass-producing automobiles, traffic travelled through central London at a sedate 20kph, with horses providing most of the motive power. One hundred years of progress later, the average speed had slowed - yes, that's right, slowed - to just 14kph. As in other major cities, traffic congestion has negated a century of determined innovation of the internal combustion engine. Figures showed that London's drivers spent around half their time in queues, incurring 2.3 minutes of delay for every kilometre they travelled.
But on February 17, 2003, London began to fight back. The British capital launched an anticongestion scheme, based on tolls, that is attracting attention from all over the world. Unlike US- and Australian-style tolls, though, there's no sitting in queues waiting to pay. Or transponders.
Instead, 688 cameras at 203 sites scattered across the 2000 hectare anticongestion area photograph the licence plates of the 250,000 cars that traverse it each day. Enter the anticongestion area marked by a red C logo painted on signs and streets. Get photographed. And get ready for a one-off charge payable for that day, irrespective of how long the vehicle is in the zone, or how many times it is photographed. At a data centre in central London, Automatic Number Plate Recognition technology is then applied to convert the photograph images to licence numbers. Motorists who don't pay the toll that day are fined about £80, automatically (see "London Traffic Scheme: How It Works", page 126).
Now, London's traffic scheme is high profile. The mayors of New York City, Paris and Tokyo are all said to be watching with interest. The reason? Fines and tolls combine to give the project a payback period of about a year and a half, and should in total generate an eye-popping £1.35 billion in 10 years - all of which is earmarked for spending to improve London's public transportation systems. Best of all, according to a report cited in The Independent newspaper in March, traffic in the city's centre had fallen by 20 per cent, improving journey times by 5 per cent. Transport for London, the UK capital's transit authority, won't report on the scheme's effects until later this month; but officials say they expect delays caused by congestion to fall by 20 per cent to 30 per cent, saving drivers 2 million to 3 million hours of frustration every year. And on present form, the authority seems set for a pleasant surprise: It had been expecting traffic to fall by only 10 per cent or 15 per cent.
In the British public sector, IT projects often fall flat on their faces. And, as these words are written, Britain's tax authorities are defending yet another botched computer-laden initiative - a new tax credit system developed for Britain's Inland Revenue department that has seen millions of working parents receive late payments, causing in some cases severe hardship.
And the omens for London's congestion charging scheme weren't auspicious: a tight implementation timetable, no pre-existing model anywhere in the world to follow, the challenge of integrating new technologies - plus a brand-new transit authority working under a brand-new mayor. The inevitable question: In an area littered with IT disasters, why should this one work? The answer: an intriguing blend of clever procurement, vigorous project management, careful design - coupled with savvy and determined political leadership.
If The West Wing ever has a traffic congestion charging plotline, expect it to look a little like what you're about to read . . .
An Uncharted Road
From the outset, the risks were obvious. Quite simply, nothing like this had been attempted in a capital city before. Nor did London's streets lend themselves to the toll-charging proposal, explains consultant Derek Turner, former head of street management at Transport for London. Cramped, narrow, convoluted and literally hundreds of years old, they would require cameras to be sited very carefully in order to achieve sufficiently high levels of number recognition accuracy.
Not only that, but the city's leadership was new, part of a process to have London voters directly elect their mayor for the first time in 2000. This led to the ascension of Mayor Ken Livingstone (see "The Man Behind the Plan", page 124). Coincidentally, Transport for London, the government agency established to upgrade the metropolitan area's public transit systems, also was new and untested, staffed by people drawn from a clutch of predecessor organisations with different policies, practices and cultures.
Of course, various anticongestion schemes had been discussed and mooted over the years. Livingstone was determined that his tenure would see action, rather than yet more talk. A 1998 independent government white paper previously had highlighted two options - a clampdown on workplace parking and a camera-based toll system. Deeming the parking clampdown to be a political bombshell, Livingstone told Transport for London to get on with the camera-based system. The deadline for delivery: February 17, 2003, one year before the next mayoral election.
For Livingstone, the political risk was huge. He could expect any failure to be viciously seized upon.
That was the bad news. But there were some positives too. Undeniably, the scheme was positive with the public. If Livingstone could pull it off, another election victory seemed assured. And why would a toll system prove popular? Quite simply, most people wouldn't be required to pay it. Although one million people entered central London by all forms of transport every morning, 85 per cent of them did so by public transport. A minority (motorists) would pay to improve the lot of the vast majority.
And there were ways of minimising the political risks. In retrospect, one of the smartest moves of the whole project was a decision by Transport for London to recognise its own limitations - limitations in terms of experience, IT ability and management time. (Officials such as Turner, don't forget, were very busy. His days were filled with bringing the capital's streets under a single management for the first time.)
Time to Outsource
Accordingly, explains Malcolm Murray-Clark, one of the two directors appointed by Turner to oversee the scheme, Transport for London made the decision to outsource critical elements of the project management of the scheme - first to consultants from PricewaterhouseCoopers, and second to consultancy Deloitte & Touche.
Very early on in the project, then, project managers identified the critical technical elements of the scheme and made judgements about the costs and risks of acquiring these as part of a big-bang solution or instead going for a best-of-breed approach. UK public-sector IT projects have often favoured the big-bang approach: The congestion charging scheme went the other way. Managers divided the project into "packages" that could, if required, be bought and managed separately.
There were five such packages: the camera technology to be used; the image management store where the images would be collected, turned into licence numbers and condensed (duplicates would occur when one vehicle was photographed by several cameras); the telecommunications links between the cameras and the image store; the customer services infrastructure, including the ability to pay by phone, Web and mail; and finally, an extensive network of retail outlets, enabling people to pay at shops, kiosks and petrol stations.
And even at this early stage, the policy of risk aversion was having an effect. Murray-Clark says one option would have been to lump the customer services infrastructure and the retail side together in a single customer-facing operation. Instead, retail was seen as a big enough challenge to be bought and managed separately. Likewise, the mere existence of the five packages underpinned a determination to go for best-of-breed in each instance - in every case, the package had to be the best available.
Putting Contractors Through Their Paces
Shortly after London's 2001 New Year celebrations had ended, Transport for London began inviting tenders for the five packages. The £71.5 million in IT-related contracts made the project large enough to require listing in the European Union's public-sector register, and tenders were open to companies throughout Europe. Again, the consultant-designed procurement strategy came into play: The camera and communications packages could be bid for separately, while the remaining three could receive bids on a combined basis, or individually. "Ideally, we wanted a single provider, but were retaining the ability to mix and match," says Murray-Clark.
While selecting the camera and communications packages was straightforward, the remaining three packages - the bulk of the IT work - proved more problematic. A long-listing process, managed by Deloitte & Touche, brought the original 40 or so bids down to four. "We went through a couple of months talking to all four, and then reduced the bidders to two," says Murray-Clark. At this point, the procurement strategy dictated another move that in retrospect seems astute.
"We told the two bidders that before we could make a decision, we wanted them to undertake a technical design study," he says. Taking three months, this step would firm up the technical issues underpinning the contracts - issues such as data throughput, how the retail channels would work, how to achieve the best number recognition performance, and what payments might be expected through each payment channel. From the perspective of the procurement strategy, explains Murray-Clark, the clever thing about the technical design stage was that it added a lot of value without actually holding anything up. "We were moving ahead but still using the competitive element of the bidding process to get the best solution," he says. The move wasn't without cost - the two bidders were each paid for the work - but the theory was that the cost of paying the losing bidder for doing the work would be more than outweighed by the benefits of using its work to improve the overall quality of the project.
It was this phase of the project that highlighted the fact that from the technical point of view, the greatest challenge was going to be the creation and management of the image store. It had to process a million records each day (picture those 250,000 vehicles moving about the city centre all day) - as well as store them for evidentiary purposes for the subsequent prosecution of nonpayers. Meeting the challenge meant carefully evaluating design considerations (such as using the most reliable technology available) and writing software code that would automatically detect which image of a passing vehicle would yield the most accurate number recognition.
From the perspective of The Capita Group, the winning bidder that emerged in December 2001, the process was a confidence booster. "The experience of the technical design study was actually quite a useful one because it created confidence in both us and Transport for London about how the scheme would work, and how long it would take," says Simon Pilling, executive director at Capita, who was in charge of the project. "The deadlines were very tight and were politically driven, and it highlighted where the risks were."
Some of those risks were outside Capita's control, of course. (London officials were anticipating some kind of legal challenge to the scheme, and the city beat back a challenge from the Westminster district arguing that it would unfairly impact them.) But for the rest of it, the company was on the firing line. The contract included a very robust set of liquidated damages against the contractor for failure to deliver against the time-scale," says Murray-Clark. "We set out the [market] stall, they agreed that that was what they would produce, we agreed to the milestones for getting there, and we agreed on the consequences of not getting there."
There was one additional ingredient: third-party oversight for the main contractor. Despite assigning people who as Pilling puts it "had spent a large part of their working lives on large government projects", Transport for London required that Capita agreed on project management methodologies with the consultants from Deloitte & Touche before work could commence. "They were the advising consultant, and their job was to report on how it was being performed," Pilling says.
Getting It Done
The project plan showed that some 300 years of effort would be required - effort that would have to be done in little over a year, if the project was to go live as planned. From the outset, Transport for London, advised by Deloitte, was determined that milestones would be rigorously monitored. "We had a statement of requirements that said: 'This will happen at this point in the process - now demonstrate it'," says Murray-Clark. "We'd decided early on that testing wasn't something that could be skimped - as problems were identified, they were ranked in order of priority, and fixed."
Capita had bid for three packages (image management, customer payments and the links to retailers), and it won them all. Transport for London, however, preferred to deal with one prime contractor, and so it awarded the company the task of managing the camera and communications parts of the project as well.
Far from complicating the task, selecting one company probably made it easier: Capita had decided at the outset that all the people working on the project - Capita personnel and subcontractors - would be physically located together in a single building in Coventry, in central England. "If you want a project to work well, it makes sense to bring all your team together - putting everybody together cuts out a lot of bureaucracy, time wasted through travelling, and means that the only videoconferencing we were doing was with the client, back in London," Pilling says.
The views of the subcontractors are hard to elicit. Essentially, Transport for London has banned them from speaking publicly about the charging scheme, a decision that insiders say is unlikely to ever be reversed. "It was a contractual decision that we made early on in the process," says Murray-Clark. "This was an extremely newsworthy project, and we needed to speak clearly and with a single voice."
But the subtext is also clear. Prime Minister Tony Blair has shown a mastery of public relations far ahead of his opponents. Livingstone wanted his administration controlling the spin on this project.
As the days after February 17 showed, the news was of course good. In fact, it's hard to find a negative aspect to report - with the sole exception that because so many drivers have been deterred from entering central London, revenue may be below expectations. But for a project with the primary objective of reducing traffic rather than raising revenue, that hardly counts as a major failure.
Why did it work out so well? In addition to the attention paid to procurement and project management, Murray-Clark and Pilling point to a few other factors. First, scope creep was vigorously guarded against - with one of the few add-ons, in fact, being the option for motorists to pay tolls through the popular SMS text messaging format. Second, Capita's deliverables were spread out over a manageable time-scale, rather than concentrated toward the project's end. And third, notes Turner: "It simply wouldn't have happened without a strong political leadership and will."
Finally, adds Murray-Clark, "The trick was to not use any new technology - all the technology was proven. What had never been done before was to integrate it - and to do it on time and on budget."
SIDEBAR: The Man Behind the Plan
London's first elected mayor made traffic congestion his main priority
The London traffic toll scheme wouldn't have happened without Mayor Ken Livingstone's commitment to it. And Livingstone's commitment wouldn't have been possible without the decision by the incoming Labour administration in 1997 to emulate an American idea: directly elected mayors in a handful of British cities, among them London. (Past London leaders were chosen from the elected city council members.) To Prime Minister Tony Blair, this looked to be a novel way of tackling perennial problems - in London's case, its poor transport infrastructure.
Blair had a nominee, but his plan derailed when a controversial Labour member of parliament named Ken Livingstone announced he would run as an independent. And run as someone who opposed the government's notion that the London Underground should be privatised. Livingstone, 58, a man for whom the words maverick and firebrand might have been invented, is a veteran London politician dubbed Red Ken for his brand of left-wing politics, such as his advocating gay rights legislation and British concessions in Northern Ireland.
In May 2000, when the dust settled on the three-way race for London mayor, it was Livingstone by a landslide.
On July 3, 2000, Transport for London was born, tasked with managing London's hitherto separately managed transport services: roads, buses, river boats and taxis. Within days, Livingstone was talking about congestion charging. One early confidante was Derek Turner, London's traffic director at the time. "He wanted to know how long such a scheme would take to implement, and what the risks were," recalls Turner. Weeks later, Turner found himself in an ideal position to find out, when Livingstone - the chairman of Transport for London - appointed him the agency's managing director of street management.
Livingstone, referring to the need "to tackle the congestion which cripples this capital city", said when the traffic charging plan began in February: "From today something is being done. If we want London to continue to be a success story for business and jobs, then we must enable people to move around the heart of London more efficiently. Congestion charging is the only option available - there is no practical alternative."
SIDEBAR: Why It All Worked
An expert identifies the reasons for the traffic project's success
Though he doesn't claim to have studied every facet of the London traffic congestion scheme, Richard Heeks, senior lecturer in IS at the University of Manchester, has seen a number of British public-sector IT projects gone wrong. Heeks, editor of Reinventing Government in the Information Age: International Practice in IT-Enabled Public Sector Reform, notes that the London traffic project's managers were guided by three principles: a narrowly focused goal, getting vendor partnerships right and testing.
"The partnership between Capita and Transport for London has clearly worked well, and worked in a way that hasn't in other big government IT projects," Heeks says.
Compared to some other government IT projects - modernising the Royal Mail or reinventing the UK Passport Service, to name two - the London traffic project "is a pretty simple and relatively focused project with very clear objectives. In contrast, many of those failing government projects have involved human elements, and they have had unclear objectives," he says.
Lastly, an emphasis on testing was crucial. "They really, really tried it out before the go-live date. They had thought about what could go wrong, tested it and sorted out any issues beforehand," says Heeks.
SIDEBAR: Vital Stats
Congestion zone is 2000 hectares (eight square miles) in central London
688 cameras at 203 locations to record licence plates
174 entrances and exit points to and from the zone
Drivers are charged when they enter the zone weekdays between 7am and 6.30pm