I was at a conference in 2010 where it was being presented. The inevitable question came up about whether this was the next public sector IT disaster waiting to happen. The answer given was no, of course. However the reason given for having confidence was that ‘agile methods were being used’. Oh dear! Anyway, the NAO report earlier this year buried that claim.
I’ve been reflecting on recent discussions on UC around one particular claim in particular. That is, the commitment to get it right, even if that means delay.
My first law of major projects is as follows: ‘Nothing derails a project faster than the unintended consequences of intuitively sensible decisions.’
Let me explain.
I was a project manager for the DHSS automation of the Unemployment Benefit Offices in 1984. This involved the roll out of systems to nearly 900 UBOs at a rate of 10 per week. Over two years there were around 30 change controls related to changes in the welfare system.
Despite the usual cynicism, most were not due to political interference/guidance, but were the results of case law, tribunals and other judicial reviews. Most were relatively trivial but one or two altered the business logic requiring changes at both ICL mainframes and Honeywell minis.
The really significant challenge of running parallel systems for two years was not in the IT, but in keeping 30,000 UBO staff trained. High staff turnover further complicated this challenge. Keeping the old and new systems in step for the roll-out period with the drip of change controls was a significant management challenge. I might say that my experience was that it was handled incredibly well by civil servants looking back at it, though it felt chaotic at the time.
Now taking six benefits and rolling them into one and running parallel systems for many years raises major concerns for me. Every mistake will get media attention, create crusades for welfare groups and flood Citizens Advice, MPs and other agencies.
As I understand it, once on UC, an individual will stay on it. Consider the following: An individual is on UC in say Manchester. Their parents are taken ill and they move to Liverpool to act as carers. Their circumstances change, but Liverpool is not yet on UC. How is the system supposed to work?
The unexpected mobility of people in Scotland played a significant part in the failure of the Poll Tax, raising costs and reducing yields.
My worry is that if UC ‘fails’, it will be represented as an ‘IT public sector failure’, when the bigger challenges, for me, lie in change and programme management.
There have been so many abortive attempts to reform welfare, so another may have serious consequences for some of the most vulnerable in society.
What would I do? I would merge one benefit into UC each year over five years starting at the new financial year and run the system nationally. In other words, timetables matter. It would make change control more manageable and reduce the costs and challenges of parallel systems. It would enable staff training to be more consistent.
I’ve been reflecting that one aspect of public sector IT projects that does differ from major private sector projects is the change control aspects of legal developments, such as tribunals and judicial reviews.
So, while ‘let’s delay and get it right’ sounds sensible, my experience is that delay makes it less likely that it is possible to get it right. Delay can increase costs, yes, but more importantly can increase complexity of deployment.
Sometimes I use an example from the newspaper industry. An outsider looking at what happens an hour before the edition is finalised, would imagine that there was chaos and little evident management. Yet, the times are very rarely missed. No-one wants to be the person who holds up the whole paper. The discipline is central to the running of a newspaper.
So which is it to be, Universal Credit or universal chaos?
