Shared Services are back on the modernising agenda. The good news first, because there is some. There is much to cheer about the internal tools used by civil servants being put in the spotlight. Amidst the digital improvements sweeping patchily yet inexorably across the UK’s public services, these tools have remained laughably poor.
For some, resetting a password on the current self-service system takes five weeks and involves three people. Thousands of civil servants are wasting thousands of hours completing incomprehensible forms; hidden millions of pounds still go wasted. And it’s not all about the money; the corrosive effects of such execrable systems on morale are insidious.
There are heartening signs, too, in the leadership lining up behind making shared services work. Matthew Coats, the SRO, is widely seen as a man going places; one would also imagine he is plenty experienced enough to give any brief that carries a career-blighting stench a very wide berth. The cabinet secretary has taken pains to personally promote the work and firmly position himself as being accountable for its success. Both have emphasised the importance of an improved user experience and applying digital standards to what’s going on inside government.
Spending precious executive time on something as esoteric as shared services sounds odd, but the alternative is to repent at leisure. The Canadian government is at this moment counting the cost of Phoenix: a seriously bungled payroll system change. Just last week, the federal government announced it was scrapping 10 years of IBM/Oracle work, leaving behind a billion-dollar bill, some very angry unions, and a backlog of 384,000 queries from public employees who are being paid incorrectly two years after the system’s launch. Nothing is rising out of those ashes for some time.
The concept of shared services has had a rather hard time in UK central government, comprised as it is of two things — sharing and services — that some parts of the Senior Civil Service lack an affinity for. Yet the basic strategic case for shared services is sound for an organisation like central government. Several employee needs are identical across different departments: booking travel, getting paid, logging holiday, and so on. Why pay for scores of systems that do the same thing?
Applying the neat theory of shared services to the messy complexity of government is incredibly hard. Indeed, so far it has proved impossible
The 2018 vintage of the government’s shared service strategy makes this simple argument. Unfortunately, it doesn’t say much else. The “Details” section of the Shared Services strategy page on GOV.UK says, in toto, “This Shared Services strategy will save the taxpayer millions of pounds by 2028.” The strategy itself — a 16-slide presentation with management consultancy fingerprints all over it — does not provide an excess of reassurance that lessons have been learned from previous attempts at building shared services.
Applying the neat theory of shared services to the messy complexity of government is incredibly hard. Indeed, so far it has proved impossible.
For it to be worth doing, you need some convergence: an agreement that the many bits of government will behave the same, and continue to do so for the duration of a contract. For all the warm words, this rarely happens, because departments generally have little incentive or desire to work in the same way as their peers.
The technology used to implement shared services is then cobbled around this complexity, at significant expense and breaking any shred of elegance it once had. Nobody is happy with the outcome. Technology becomes the scapegoat for this mess, but the root cause lies in deeper truths about institutional rules around accountability.
Rather cutely, despite the mention of savings, the current strategy contains no numbers at all. Lessons have obviously been learned from the last time the Cabinet Office tried a shared services strategy. The last effort, published in 2013, publicly promised between £400m and £600m in savings. A 2016 National Audit Office report on the programme reported a return of £90m in savings, from an investment cost of £94m. In his blog, Coats described these as “major savings”, which is certainly one way of putting it. I’m sure the NAO will be watching the new initiative very closely.
The opportunity of shared services lies in finding out about employee needs in an internet-era civil service, and delivering something that offers the flexibility to address those needs as they change over time
I suspect I’m not alone in feeling a little wearied by the idea of shared services. At this very minute, there is at least one government department in the process of implementing a shared IT system where one consequence has been that every official is writing several weeks’ worth of diaries on paper. There’s a bug in the Microsoft calendar migration, apparently. This kind of thing is a double blow for civil servants, because not even the most cynical actually wants new systems to disappoint. This makes it all the more maddening when they invariably do.
But that need not be the case. There is a huge opportunity for the civil service in getting shared services right. One of the biggest lessons from designing digital public services was that it is not enough to simply digitise the existing process. Shiny front ends are not the answer; though in this case, even some of that tinsel would be welcomed with open arms. A slightly better look and feel to HR and finance is opium for the masses though; serving to conceal tangled lines of accountability lying just beneath.
Fixing that root cause will take decades, and quite possibly a fairly major war. Being more realistic, the opportunity of shared services lies in finding out about employee needs in an internet-era civil service, and delivering something that offers the flexibility to address those needs as they change over time.
The primary aim of the shared services programme should be to make the failure of new employee services cheap, fast and reparable, rather than the expensive, drawn-out embarrassments many civil servants have got used to. If it can do that, while iteratively improving those services based on regular user research and avoiding lock-in to lengthy supplier relationships, massive incremental improvements are possible. Punting out some 10-year contracts to SAP or Oracle and crossing our fingers, meanwhile, is a one-way ticket to Phoenix. We’ve been there before.