While we should take our hats off to Iain Duncan Smith for his thorough analysis of social welfare problems, he is being misled when it comes to delivering the solution, the Universal Credit (UC). Anyone claiming for the new Universal Credit will be obliged to use the internet; in the words of the designers the service will be “digital by default”
This is the first of two fundamental weaknesses in the plan for delivering the UC: it will require successful building of a large-scale computer system. As Gauld and Goldfinch (â€˜Dangerous Enthusiasmsâ€™, Otago Press) point out, in a veritable tsunami of evidence, the vast majority of large-scale computer systems fail.
Indeed we have seen reports in the Daily Telegraph this week that the Universal Credit has already been moved to the top of George Osborneâ€™s warning list of projects that could fail. The chairman of the Commons public accounts committee has called the plan â€œa train crash waiting to happenâ€. Unsurprisingly, the concern centres on the IT.
The Child Poverty Action Group is not confident either. Chief executive Alison Garnham told Public Servant: “The Department for Work and Pensions is not famous for its success with new IT projects, and I really worry, because everything goes into one systemâ€ and “If it fails it will have really serious consequences.”
Duncan Smith has been persuaded by the IT companies this wonâ€™t fail as other projects have (think NHS patient records system, regional fire control centres and passport office). But they would say that wouldnâ€™t they? The civil servants are confident too. According to a senior DWP source quoted in the Daily Telegraph, the department is confident it can deliver its side of the Universal Credit system on time. IT companies have persuaded Duncan Smith that this time things will be OK because they are doing â€˜Agileâ€™.
Agile â€“ the building of computer systems in an iterative, â€˜liveâ€™ way â€“ is merely to do the wrong thing faster. Which brings me to the second fundamental weakness: If the computer system actually gets built, the service will fail, because computers are terrible at absorbing variety, and taxation, credits and benefits are high-variety problems. Using computers to deliver the service will amount to an attempt to codify, in rules, the eligibility and entitlements for claimants; and as rules can never deal with variety, the consequence will be poor-quality, hard-to-get services for those who are the most vulnerable in society.
The plan is driven by an obsession with cost. Paradoxically, it will drive costs up. The assumption is that web-based transactions will be cheaper. But this is to confuse transaction costs (which will be lower on the web) with the true costs of service â€“ the total number of transactions it takes for citizens to get a service. The failure of the web-based service to resolve peoplesâ€™ problems will generate massive amounts of what I call â€˜failure demandâ€™ â€“ demand caused by a failure to do something or do something right for the customer.
I introduced Terry Moran, the DWP man leading the UC project, to my clients, in order for them to explain what they have learned about computer systemsâ€™ inability to deal with high-variety services. In addition there is clear evidence that local fraud detection is far more effective than nationally-organised fraud detection. More importantly, I explained to Mr Moran how local authorities that have rejected DWP guidance on how to manage their housing benefits services are delivering benefits in a matter of days, dealing with each claimant in the round, preventing further transactions with other services. These profoundly better and cheaper services â€“ saving 20 to 40% of operating costs â€“ illustrate how peoplesâ€™ lives are complex and their needs are better served by people, not computers; only people are able to absorb the variety of claimant needs.
I explained to Mr Moran that the same principles could be used by local authorities to deliver the Universal Credit; they would be able to create a Universal Credit service in months rather than years, with no major investment in information technology. It would be a service that is based on need and would help claimants in the context of their communities, building relationships that will be constructive aids to achieving the purpose as set out by Iain Duncan Smith â€“ helping people achieve a productive life.
But the train has left the station. In the words of the chairman of the Commons public accounts committee the plan to deliver this brilliant idea is a train crash waiting to happen. It is guaranteed.
Professor John Seddon is an occupational psychologist, MD of Vanguard Consulting and visiting professor at 3 UK universities. He has received academic honours for his contribution to management science.