Why I believe the government has it wrong on IT

In recent weeks, as the main parties look for somewhere concrete to cut costs in order to locate the billions they need to save, they have alighted on IT as as a strong contender for saving money. And who can blame them?

Government's track record on IT is appalling. Huge overspends on systems that fail to deliver, with just a handful of large suppliers carefully tying up the market to their benefit, and decidedly NOT to the advantage of both government and tax payer.

The Child Support Agency's CS2 system, for example, is riddled with "insurmountable bugs" that mean many cases have to be handled manually: 19,000 in 2006 rising steadily to 75,000 by September 2009.

The problem, here, I believe, is how government buys its IT. It appears to completely lack the expertise to make procurement decisions that will deliver the required outcome. I have met a lot of intelligent people in local government who are quite incredibly committed to delivering good service. They know what they want from IT but seem quite unable to get it from the handful of providers who have found seats on this particular gravy train.

Recently we tendered for some work from a local authority, which had talked about developing an open source system that could be shared with other councils. Yet, when we received the paperwork, it was clear that the current provider had the authority locked into its own CRM solution. Elsewhere, when we have tendered successfully, we have delivered working solutions at a fraction of the price the larger software houses have quoted, in once case coming in at less than a sixth of the cost.

So, as I said, I can see why government has had enough of IT, spending billions on projects that are either abandoned or, once delivered, don't meet the requirements. However, I think it would be a serious mistake to abandon IT solutions. Going back to the CSA for a moment, the National Audit Office has calculated that when a case can be processed through the system, the cost to the tax payer is £312. When processed manually, the cost trebles to £967. The 75,000 cases being processed manually last September were costing seventy-two and a half million pounds, instead of twenty-three and a half million, if CS2 had been able to handle them, a difference of nearly £50M.

I believe, then, that the government does need IT. Furthermore, if civil service numbers are going to be allowed to decline by 40,000 jobs (and probable a lot more), then I would argue that government needs IT more than ever. Civil servants need to be freed up from manual processes and administration, to focus on serving the public. (See my blog earlier this month, "Good business systems are about liberating your staff, not getting rid of them.")

However, the solution to the government's problem is not so difficult. Let's look at the issues for a moment, which primarily arise from the simple fact that large projects are notoriously difficult to handle.
The longer the project, the more legislative changes it will suffer over its development cycle.
Big projects are hard to manage, where large teams and deliverables need to be coordinated.
The huge amount of analysis that is necessary means that detail is skipped, so requirements are misinterpreted or lost and parts of the solution don't integrate.
Vague specifications lead to inaccurate costs, so the supplier ups the price to protect themselves. (And add to this the fact that large companies have high salary overheads, so they want as many of their consultants as possible on any project.)

Projects away from the public sector are far less likely to be run this way, with the client wanting far more clarity about precisely what will be delivered, as well as how and when, before committing any money to the project. Government should learn from this, commissioning comprehensive, precise specifications as a separate part of the project and hiring third party companies to evaluate them. Projects should be broken down into manageable releases, such that everyone involved has a clear grasp of what is going to be delivered, and on what date and at what cost. What I am saying, in effect, is that the solution is to turn these large projects into a number of smaller projects.

You might, quite reasonably, argue that as someone who runs a software house, I have a vested interest in this argument and that may be true, but I believe the figures speak for themselves. Time and again, I have seen software benefiting my clients' companies, including some local authorities. Government must not turn its back on IT; it will be more vital than ever over the next few years. What it does need to do is to learn how to buy IT, including being a bit more discerning about those suppliers from whom it buys.