Last turn of the ID cards
Education, education, education. The Government promised it and never mind the cost. By golly, they're delivering. With a cost-to-date probably exceeding £100m, the ID cards scheme is proving one of the most enlightening examples of IT project management ever to grace the public sectors — and that's before a single line of code has been written. For everyone who's wondered how to control a large project, here is the answer: see what's happening with ID cards, and do exactly the opposite.
Know why the project exists. After cycling through terrorism, identity theft, benefit fraud and commercial efficiency, the only consistent purpose for the ID card scheme appears to be to issue ID cards. Confusing the purpose of a project with the project itself is almost irresistible without strong, clear, consistent guidance: it is also a guarantee of failure. Pretending that a project is intended to do one job while really fulfilling another is the stuff of spy novels and science fiction.
Have a cast-iron business case and be ruthless about keeping to it. You're going to be spending a lot of money. How much will you get back, and how? Miss this one, and you'll be easy meat for the next predator with a spreadsheet. In the case of ID cards, no business case was made or promised — which killed at birth the spirited refutations of the numbers other people produced.
Listen to the users. In almost all cases, the people who pay your wages are the ones who'll use or be affected by your project. That same channel can also deliver dismay, revolt and disaster. Outside a police state, you can't force people to use things they don't want to use and which they perceive as harmful. That goes for CRM just as much as ID cards. Miss out the user, and you're missing out on an interface with reality.
Specify the project and be ruthless about its scope. You know what the project will do, you know how much it will cost, now here's how it's going to work. As important as specifying the core of the systems is deciding what potential features will be excluded. No IT project is without room for near-infinite expansion: no IT project has the luxury of including more than a fraction of those ideas. Leave room for more later by all means — and specify those interfaces — but keep it clean and basic at first. In the case of ID cards, nobody knows what's in the database and what's out, where the cards will be used, or who can see what when.
Ignore any of the above, and you're in trouble. Ignore them all, and you'll be spell-checking catastrophe on your CV for years to come. We hope that as he surveys the mess he's inherited, John Reid will take the opportunity to employ the last and best rule of project management. When you find yourself in a hole, stop digging. When we see that happen, it will be an education for all of us.