IT procurement in government – and, most likely, the public sector as a whole – has for many years been a fast-growing industry. In my experience, around a third of all client-side IT staff (those serving the commissioning and operational needs of other staff) have been engaged, directly or indirectly, with procurement. A similar-sized army has been employed by the supplier side of the service. Reprocurement of outsourced services typically takes years: in 2011, for example, one smallish agency was already starting activity to reprocure on a contract that ran till 2015. But it’s not just the initial procurement. Because of endless changes and often meaningless service level agreements, the pace never slackens; procurement teams are kept very busy.
So why does this approach need to change? Well, the reasons are both obvious and legion. Guesses on the number of client-side IT staff in the public sector run about 60-100,000. If just 25-30 per cent of them are involved in procurement, at an all-up cost average of £30-40,000, the annual cost of procurement is £450m-1.2bn.
For such a large business, you may think this is worthwhile – but public sector workers themselves don’t see value for money. The service provided has, in the eyes of most end users (public service workers and citizens), been woeful. It has been expensive, with costs to change one line of code reaching £50,000. It has been slow: it often takes 12 weeks to commission a server. It has been unreliable: ask any public sector worker.
Of course, not all services are like this – there have been some exceptions where smart people, working hard, have kept things going well. But the system works relentlessly against these heroes.
Apart from cost and quality, the third disastrous leg of our approach has been duplication. Often, though organisation A is spending thousands on procuring, testing and accrediting a service, organisation B nonetheless undertakes exactly the same activity, to exactly the same standards, and often at exactly the same time. It’s lunacy.
There is another way, and it doesn't require endless battles with the EU or security services. When we developed the G-Cloud framework and approach, it became clear that the UK had often chosen to interpret EU legislation in a draconian way that served to slow procurements down without improving the outcome. In Germany, procurements were typically twice as fast as ours.
As we worked through realising our vision, it became clear it was possible to do things so much better. First, we could complete the universal procurement components – those elements that are identical across the public sector, often comprising 80-90 per cent of the process – and make the results available, ensuring that all other public organisations could reuse them; this would save public bodies lots of time. Second, by security-accrediting services – where appropriate – through a pan-government assessor, we could take up to 90 per cent of the cost out of accrediting these services for other public bodies. Third, making all services visible and comparable in price and functionality would break us out from the world of secrecy in which one organisation couldn’t tell another what they’d paid for a service. This transparency, along with the relative ease of getting on the G-Cloud framework, opened the door to hundreds of SMEs and emerging suppliers – further driving down price and improving service and innova
tion. This approach is also being considered now by other countries and for other commodity products.
Getting this introduced across the public sector’s thousands of organisations and tens of thousands of procurement staff is not easy. It will take a while to get this approach universally accepted, but the signs are good: last week, the Department of Work & Pensions and Home Office declared, in effect, a ‘G-Cloud first’ approach. Others will follow.
Unfortunately, the Cabinet Office has not helped. The G-Cloud programme remains woefully under-resourced, because of an unrefined macho approach to staffing levels and recently increased levels of bureaucracy. Yet organisations will need support to make the change. I am sure that Cabinet Office minister Francis Maude and his non-execs can't be aware of these problems: they were vociferous in their support when head of government IT Andy Nelson and I presented to them in the summer of 2011.
Support for the old ways is still strong – and for no reason other than the comfort of avoiding change. So the government ends up giving existing suppliers a gentle squeeze, and securing a modest discount rather than the transformational change that would be possible. Five or ten years ago, savings of 10 per cent would have been good, but today 90 per cent is becoming eminently achievable.
The work done by Government Digital Service (GDS) is a classic example of how to get it right: it’s an organisation focussed firmly on customer need, building an in-house (where appropriate) team that uses modern, low-cost, mostly open source, often cloud-based solutions to deliver vastly improved services at dramatically reduced cost, at speed. But where is the GDS equivalent in the world of shared back office services? The model used by the GDS team is equally applicable here. Yet Cabinet Office support for that model in shared services is invisible. Instead, we have a solution from the 1990s that will achieve ten per cent savings – and that money, I guarantee, will be lost in an increase in change management costs.
See Cabinet Office response
Chris Chant is the former head of the government's G-Cloud programme. He retired last April.