Monday, March 17, 2003
Why do government IT projects fail?
Projects fail for pretty much the same reasons all over, whether they're private sector or public sector. The top few are all about stakeholder engagement, decent business case, focused risk management, capable people and day to day project management (a la Fred Brooks). In the public sector there's a special reason though. This week the press picked up on a report in Computing noting that over £1.5 billion had been "wasted" in government IT projects since 1997, although a big chunk of that came from just a couple of projects. If you were able to add up the 1s, 10s and 100s of millions here and there across projects, doubtless the number of billions would rapidly hit double digits. And I have no doubt that if you were to do the same in Citibank (where I've worked), Cable and Wireless (where I've worked) or ICI (where I've worked) the numbers would be not dissimilar in %age of IT budget terms. Simon Moores followed up the Computing piece, in his Computer Weekly column, with the observation: "Has anyone been sacked or lost their index-linked pension? I think not somehow. If any IT Manager blew a million away in the private sector – unless perhaps he worked for AOL -Time Warner – he’d be toast but in the public sector, it’s possible to sleep at night, knowing that £100 million or so of one’s responsibility is about to go ‘Tits-up’." I know the FT followed up on the Computing piece, but I haven't yet subscribed to the online version. Also, I have a page from Computer Weekly ripped out in my notebook right now that I've been carrying around for a couple of weeks or more. It observes that IT departments should scrap at least a quarter of current running projects to improve success rates. So if government IT failures are hitting 54% (from the HMT Green Book) and we scrap a quarter, that means we'd have at least a 66% failure rate (because we'd certainly wrongly identify the projects and cut the good ones). I have another piece from Computer Weekly in my pile to read that is headlined "EffortStop Duplicating ". There it's noted that "companies are wasting millions of pounds because work is being repeated and expertise is not being shared". And, remember, that's corporates, not private sector. And the special reason why public sector IT fails? Delivery is not a valued skill in government. There is no "fast stream" route to the top for delivery people - find me a permanent secretary (a far more vital position than it sounds, for you US readers, this is someone who is at the very top of the civil service tree) who has actually delivered real projects - IT or otherwise. For all their skills, the perm secs have not done that, nor have the deputy perm secs or anyone else at the top of the tree. So, when you're young and aspiring to the top, you don't go into a delivery route because there is no way up. Glass ceiling? More like rock solid cave ceiling. There are, as always, exceptions. Senior people in the Revenue (and you shouldn't believe everything you read in the press about them) are working hard and delivering results, folks like Richard Granger prove that you can come in from the outside and do well - but the exceptions are few. So what happens in government is that important, even vital, strategic projects are overseen by policy wonks. And I don't mean that in a derogatory fashion - techies are geeks, delivery people are mavericks and policy people are wonks. Their focus is on process, documentation and doing things the right way - certainly not in doing things right. The OGC's Gate reviews are trying, with some success, to address this but ask them what the top 10 reasons for failure were when they started the reviews and ask what the top 10 reasons are now. Any change? I don't know for sure, but I doubt it. So when the policy folks are working on process, the delivery people are gnashing their teeth wondering when they get their chance to shine. If there are delivery people. By the time the few that are there get the project to hand, it's over-scoped, behind on time and absent of money to do things right. And because government puts policy wonks onto projects, suppliers are forced to do much the same thing. Woe betide the supplier who mixes up the relationship by putting someone creative and delivery focused up against a policy person. That's a bad idea. It doesn't work. So, for projects to work in government, the delivery people have to become valued, given a route to the top (if that's what they want). At least they should be given the project to run before it's too late. Then, suppliers will be encouraged to put delivery people in their teams to match. The combination will prove powerful as people with common aims - to make a difference through delivery - fuse and work to make the agenda happen. But if it's too late, the project never arrives, arrives under scope, disappoints the stakeholders and fails to ignite the transformation of government that we all know is possible - the one that allows people to get what they want from government, whether it's through .gov.uk, .co.uk, .com, the phone or whatever.
Posted by Alan at Monday, March 17, 2003