Delivering complex IT projects is never easy, but at National Lottery operator Camelot the pressure to do so is exceptionally intense.
When it undertook a major project to replace all of its retail terminals, Camelot was facing fines of £1m for every day the lottery was unavailable - fines it has never incurred.
Moreover, as part of its winning bid to run the lottery for a second term, Camelot undertook a major programme of introducing new products and channels for placing bets of play, which required comprehensive new IT.
Tim Newing, Camelot's IT director, said, "We said that we would completely refresh the technology that had run the National Lottery for the first licence and introduce new channels to market, such as the internet, interactive TV and mobile phones, and all while preserving the integrity of the existing business."
The first step in managing such complex IT projects is to understand what complexity actually is, said Newing.
"A complex IT system or project is one where multiple elements need to come together for success. Perhaps it is lots of different work streams that are contributing different things or maybe there are lots of different people involved, such as contributors, stakeholders or recipients.
"Other factors can make systems and projects difficult, but they do not necessarily make them complex.
"Common difficulties include shifting requirements, business changes that affect the IT business case, challenging dates or a technically difficult or leading-edge system.
"But these difficulties can also arise in fairly small-scale projects, just as they can in a complex system. What makes a system complex is its multi-faceted nature."
Newing has three golden rules when it comes to getting complex IT right. First, do not think of it as a complex project, but as a collection of more simple solutions.
"This will make the entire project more manageable, as you make a series of changes over a defined time period," said Newing.
"For big projects this approach will mean delivering more with a 'little and often' approach rather than a 'big bang'. For established systems, it will help with issue management.
"So, to overhaul a company's systems you would introduce a succession of small-scale improvements to gradually establish the full new system. This would be done within a defined time period - say, two years. Each individual system would need to interface with the existing or outgoing systems as well as future ones as they are introduced.
"It might be slightly more work overall, and you need to be mindful of the different interfaces you are creating, but there is a massive advantage in that any errors that occur are small scale because the changes you are introducing are incremental."
Newing took this approach when migrating to Camelot's new look, second-licence IT infrastructure,
"Our approach to this daunting challenge was to break it down into more manageable elements. We began by concentrating on converting what we already had, leaving the introduction of new channels for later."
A number of separate activities were started all at once, splitting up the different work streams according to functions.
Newing said, "We ordered new lottery retail terminals with additional hardware so they could connect to the old network, and we wrote a special application so they operated in a very similar way to the previous ones.
"This meant that when we came to roll them out we were not dependent on other parts of the project. By taking this approach we not only minimised disruption to retailers, but we also avoided penalties from our regulator.
"We set up new equipment in our datacentres and ported the existing applications to the new platforms, separating that element from the functional changes we went on to make. The new equipment could connect to both old and new terminals.
"We also started to design our new network and went through a process of proof of concept, development and lab testing to design a solution over ISDN that gave us the performance and tariffing that we needed.
"When we came to roll this out we had to add hardware and software to the new terminals because of our separate approach.
"But it de-risked the entire project, and today we have the largest private, dedicated ISDN network in Europe, servicing more than 26,000 terminals with high availability and good diagnostic capability.
"Only when that roll-out was behind us did we then turn our attention to launching the new interactive channels that we had committed to, and refreshing many of our back office systems."
Newing's second golden rule for managing complex projects is to manage outside the "business as usual" structure.
"We set up a separate business unit - the interactive division - which was solely focused on developing and launching the new channels," he said.
"The team was drawn from all areas of the business, from IT through to marketing and finance. This was an essential element of its success.
"The importance of a multi-disciplinary approach, together with co-location and a clear single focus, cannot be overestimated."
Again, the task was split into manageable stages. "We started by launching instant win games - an online version of scratch cards - on the internet in February 2003," he said.
"This gave us the opportunity to make sure that our systems, operations and business practices were robust with a small-scale business before launching any of our bigger, established games on the web."
By December that year Camelot was ready for a period of major expansion. The main game, Lotto, was launched on the internet that month, followed over the next 12 months by all the other offerings being made available on channels such as the web, digital television and mobile phones.
"This approach - setting up a separate business division, launching services a bit at a time - proved to be a big success," said Newing.
"We went from taking our first tentative steps into online retailing to today's more than 1.8 million registered players and sales of £4m a week."
However, at some point, the new systems needed to be integrated into normal operations.
"With the systems successfully developed and launched, the final stage in the process was to integrate the special interactive unit back into the main business," said Newing.
"This was done in April 2005. The process was not always smooth, but it was a success. Interactive sales now sit in the main operations directorate at Camelot, raising over £1m a week for good causes."
Newing's third rule for complex project success is to make sure people have a really good reason to complete the project.
"This is not just a matter of having a plan with stretching but realistic targets. It is vital to ensure that all the loose ends or extra phases and any 'nice to haves' don't get a life of their own," said Newing.
"You need to avoid the project, or bits of it, carrying on after completion, and upsetting your budgets and business plans."
This particularly applied to the National Lottery Fast Pay system, developed with Tesco, whereby lottery tickets can be bought at supermarket checkouts.
"Many aspects of this system break new ground. Lottery tickets are printed on Tesco checkout printers, we use Tesco's own network for the transaction, and Tesco is sharing its Epos terminals with us, which are therefore running a Camelot application."
Despite splitting the development into phases, each with a long period of testing, followed by a small-scale pilot and a controlled roll out, "as things progressed and we learnt more, it became clear there were opportunities to enhance the project - and then enhance it some more," said Newing.
"We could carry on finessing the system, and perhaps never stop."
But the temptation to continue tweaking the software indefinitely must be resisted, said Newing.
"Making improvements to usability once the software is deployed is fine, providing you apply proper governance and control over any additional work.
"But you need to have the discipline to know when and how to stop. If you do not, it will cost you money - probably lots of money.
"You need a cut-off point, a definitive deadline. With Fast Pay we had one - the team had to move on, at a fixed date, to start developing another system."
CV: Tim Newing
Tim Newing has been IT director of Camelot since 2002. Prior to that he was head of projects and networks at the lottery operator, after joining the firm from IT supplier GTech UK, where he was technical manager. Newing is to leave Camelot in August to become IT director of EasyJet.
Golden rules for complex IT project management
Newing has three golden rules for getting complex IT right:
- Ensure that you are thinking about multiple small projects rather than a single complex one, so that they can be addressed with simple solutions and phased in, in terms of both time and functionality.
- Separate complex projects from the day-to-day business.
- Plan the closedown carefully. Make sure that people have a really good reason to complete the project so that bits of it do not carry on after completion and upset your budgets and business plans.
Read article: How English Heritage manages its outsourcing
Vote for your IT greats
Who have been the most influential people in IT in the past 40 years? The greatest organisations? The best hardware and software technologies? As part of Computer Weekly’s 40th anniversary celebrations, we are asking our readers who and what has really made a difference?
Vote now at: www.computerweekly.com/ITgreats