Last week I spent a day at a public-sector shared services conference in Manchester. The experience moved me to write… I hope it moves readers to do all they can to alert public-sector leaders, politicians and anyone who ought to care to what has been, and will continue to be, an ill-conceived folly – a shocking waste of public-sector funds.
As I expressed it in my title slide: Does sharing services ‘bake bread’? What does? More on what I had to say later.
The opening keynote was from the interim leader of the Government Shared Service programme (‘interim’ – there’s a clue); a very nice man. If I’d been playing Consultant-speak Bingo I’d have had a full house in no time. We heard:
We aim to be the best civil service in the world, we are already world class
Shared services is a lever for change
Shared services in government is no longer a ‘programme’ it is ‘strategic’
We are to be a centre of excellence
We are re-shaping our governance to be more inclusive and more strategic
Technology drives a different way of doing things
The wider changes in society mean greater expectations amongst the public
Implementation will be Agile
We are going to professionalise our people
We will drive value and efficiency
We will take advantage of off-shoring
We want more self-service
We will be using robotics and artificial intelligence
I’ll come to my misgivings about much of that when I get to what I had to say when it was my turn. Interim Leader described how this was going to be done: A decision has been made to split up the management of technology and management of the service centres as this, we were told, will drive lower costs. We heard about there being three technology platforms, there will also be ‘process convergence’; process standardisation will be ‘relentlessly’ driven and there will be ‘standardised functions across government’. The services will be ‘designed around user needs’ and will be ‘intuitive, easy to use and mobile’. Just as services will be digital, government will be ‘digital on the inside’. And everything will be done in a ‘collegiate and collaborative’ way, establishing a ‘cohesive community of practice’.
What’s your level of confidence? Would you invest?
Well, actually, you are investing.
Interim Leader assured us they had already saved 20% of their operating costs. This was a surprise to me. The most recent National Audit Office (NAO) report said the savings were less than the costs. I asked Cabinet Office attendees, they told me this 20% was the discount on current operating costs that is the basis for the contract with the private-sector provider. So that’s a promise. And we know from the NAO that there are hundreds of ‘change orders’ in the mix – these are the means by which providers can get more money from their entrapped clients; ‘oh, that’s beyond the contract’… I remember being told by an IT consultant that IT contracts ensured the client got ‘breakfast lunch and dinner’… but ask for anything else and the charges can make your pips squeak. I also learned that the Cabinet Office has a benchmarking cost-of-transaction activity. If you are a regular reader you will know that the cost of transactions tells you nothing about the cost of a service.
So there I sat, feeling very animated, even furious. Perhaps I shouldn’t have been surprised after all this is nothing new. But that’s my point; given the exorbitant cost of well-publicised failures of shared services shouldn’t we be just a tad more cautious?
The NAO has published plenty on the costs of failure and their representative was up next. Interim Leader had been generous in his praise of the NAO, even though the reports were critical, and the compliment was returned; NAO Representative declared himself to be ‘supportive’ of Interim Leader. Warm feelings were shared. I didn’t feel so cosy.
NAO Representative would have ticked further boxes on my Bingo card. He told us that shared services should be seen in conjunction with ‘transformation’. By this he meant the quality of business cases – they should be ‘realistic’, represent ‘value for money’ have ‘clarity’ and show understanding of ‘risks’. I’m sat there thinking this juggernaut is one humungous fat risk. Business cases are dreams.
Transformation also means ‘leadership’, ‘governance’, ‘operational planning’, data quality and so on – my mind wandered before I could copy the rest down and the slides moved on… Apparently there has been too much turnover of leaders. I turned to see how Interim Leader reacted to that. It came to mind that the blessed Universal Credit fiasco has churned leaders, some sacked, some resigned, some sick. Who’d want the job? Why did Interim Leader look so relaxed, confident and kindly?
NAO Representative showed us a slide extolling anticipated savings, more promises; no doubt the numbers taken from business plans. He joined Interim Leader’s theme in discussing the pros and cons of seeking collaboration versus mandating involvement. You may recall Frances Maude, previous Minister for the Cabinet Office, insisting, through frustration, on mandating. After all, the plans demand high volumes of work moving to shared services.
Despite the series of NAO publications evidencing the expensive failures of shared services we had no explanation other than his general remarks on what’s important in ‘transformation’. Was this what was missing? How confident can we be? Is this trying harder? Is it doing the wrong thing righter?
My mind wandered back to a meeting at the NAO a few years ago when the failures first hit the news. I had always regarded the NAO as an important independent voice; I was invited and was pleased to attend. The purpose was to discuss the reasons for failure. I told the meeting the things I told this conference (see later); and in much more detail as we had a whole day to discuss. But nothing of what I’d said got into the report of the meeting. Every other participant came from IT and major consultancy firms. And they, like NAO Representative on this day, thought it merely a matter of doing it right.
A quick search on the internet revealed NAO Representative used to work for McKinsey, I knew already his boss used to work for PWC. Forgive me for speculating that the NAO, like many government departments, has been hijacked by the Big Consultancies.
Next up was IT Man. More boxes to be ticked on the Bingo card. Digital and Artificial Intelligence (AI) are, he proclaimed, ‘disruptive’ technologies. ‘You don’t want to go the way of Blockbuster’ (yawn, how often has that been cited as the ‘killer’ case?). ‘It’s a mobile / social media world out there now’; ‘the technology is coming to the office’; ‘traditional organisations are ‘edging towards digital’, the two ‘operating models’ will meet, ‘AI will connect them’. It’s all about doing things in an ‘Agile’ way.
And he banged the same drum that is the basic rationale: transactions will be cheaper; we will save enormous sums of money. Oh really?
When my turn came I put up a slide with the politicians’ view of shared services. A report from the Public Accounts Select Committee had this to say:
“Central government has long pursued shared service centres as a way to reduce costs and free up resources from back-office functions to provide better front-line services. The principles of reducing costs through using shared services are straightforward and widely understood, combining two key elements: one element is to standardise processes and services so that they can be provided in a consistent and repeatable way, in high volumes, by a single provider on a common operating platform; the other element is to outsource operations to an organisation that can specialise in providing a service and, through economies of scale, can offer the service at a lower cost.”
This is the gist of what I had to say:
These are politicians. What can we expect them to know about management? How did they come to this view? By listening to advisers; and amongst those were the IT and Big consultancies – in the early days there was an all-party group on shared services, serviced by the National Outsourcing Association whose members were… yes, you guessed.
Rather than being ‘straightforward and widely understood’ it would be more accurate to say the idea is plausible and widely embraced.
Rather than having ‘two elements’ as described above, there are four basic arguments. The first is ‘a common operating platform’; a euphemism for a large-scale IT system. What politicians ought to know is that 90% of large-scale IT systems fail. 30% fail completely (e.g. Single Farm Payment, NHS patient records and others) 60% fail in as much as they cost much more than planned, are not fit for purpose and don’t deliver the planned savings; lots of those.
The second is standardisation. Standardising service work drives costs up. Simply because a defining feature of many service organisations is the variety of customer demand. If you push high-variety demand into standardised processes you drive costs up.
The third: economies of scale. There are two arguments for savings from economies of scale in service organisations: less of a common resource and lower transaction costs. Less of a common resource, which means fewer managers, IT systems, buildings and so on, is a sound argument, but not always easy to achieve. For example sometimes it is hard to dispose of buildings or existing IT contracts; an argument that purchasing one IT system between three organisations that are sharing saves the cost of two is hardly an argument for savings and, in any event, was an IT system necessary? Regardless, any saving from less of a common resource is relatively trivial. Trivial though it is, it is the only real evidence the protagonists can cite. The much larger savings are promised to come from lower transaction costs; this is always the dominant argument in the business plans. But, as regular readers will know, if you focus on cost your costs go up.
And the fourth argument is that services should be outsourced to specialist organisations. When we do that we lock in cost escalation.
I asked how politicians react to the abundant evidence of failure. Admittedly they’d be unaware of failures in the private-sector, and there have been many. But public-sector failures have been in the news regularly. Think Cornwall, Somerset, Sandwell, Liverpool, Rotherham, Birmingham, Suffolk, Account NI, Research Councils, Centralised police call-handling in Scotland among many others. And news of the same in Australia, New Zealand and Canada where governments have imagined the UK government knows what it’s doing. And the only reaction I have witnessed is the same as we heard from Interim Leader and NAO Representative: it’s OK if you do it right, so let’s improve the governance and project management.
I said there are three types of shared services out there: those with true but minor savings from less of a common resource (e.g. local authorities), those with no knowledge of savings as there was no original cost baseline and which, if you talk to the users, don’t work too well and those that have failed or are in the process of failing; the categories not being mutually exclusive. So why do they fail?
The simple answer is they are classic command-and-control designs. To illustrate what’s gone wrong I told the story of a private-sector client that had, when we first met, had outsourced its service work to India. I named the client – a household name – and I shall name them in my next book. Essentially the story was they had decided to repatriate the work to the UK. They had 500 people working in India and 200 in the UK. They had a plan. Quite naturally the plan was to hire 500 people in the UK to take this returning work.
We suggested they put the plan away and instead get out and study what was going on. They had a classic industrial design, exactly the same as we see in shared services: front-office / back-office, SLAs, standard times, activity management, specialisation, standardisation and so on. These systems create enormous amounts of what I call failure demand. Seeing that for themselves energises leaders. At first they think the causes are people not doing as they should and/or processes not properly designed, but they learn through studying that the causes are everything that they think of as good management (the features above); failure demand is systemic, a natural consequence of an industrial design, and it starts to dawn on the leaders that their obsession with managing cost is actually creating costs. And that’s the big problem with shared services, they are industrial designs.
I outlined the better design that the leaders put in place. Thorough knowledge of customer demand dictated the expertise required to give customers what customers needed. No longer were service agents constrained by activity management instead the measures of performance were defined in customer terms; I describe this as designing against demand. It is to focus on effectiveness not efficiency. The result was all work was returned to the UK, the quality of customer service jumped and it only required 300 people. Who, I asked, would have put that number in a plan? Designing for effectiveness drives costs out of a service by increasing capacity; failure demand drops rapidly.
It is the central point. Improvements of this magnitude have been evidenced in the public sector too, and they are the result of changing the way we think about management. The shared services strategy reinforces everything that is dysfunctional about our current style of management. Think back to the Blair target for every local authority to have a call centre. When they obliged, they all discovered they had more demand than they had anticipated in the plan, just as had occurred in the private sector; the design was creating failure demand. It was chasing efficiency – lower transaction costs – the wrong thing to focus on.
I gave examples of the enormous savings local authorities have achieved by studying and using that knowledge to design for effectiveness. In people-centred services many lives are quickly back on the rails, services are vastly improved, costs fall dramatically. I pointed out that if they’d shared their service (e.g. share a call centre, share a building in the name of ‘integration’) they’d have got nothing of the kind but would only have cemented in conventional command-and-control management practices – the very things that had been revealed as dysfunctional when studying.
I talked about the housing association I mentioned in my December newsletter that had dumped a half-million-pound IT system, which was two years into implementation, when they studied and redesigned their repairs service. Following study and redesign, and having a design that drove costs away, they knew what they needed from IT and built it for £40,000. My point was simple: IT is the last thing to do, not the first.
These examples, I didn’t have time for more, show the scope for genuine improvement of public services and they operate at costs that are so much lower than anything that would have been put in a plan.
Finally I gave advice: If you are thinking about sharing services, don’t. Instead do as these examples did, study the services as systems, redesign them and if you need IT make it the last thing you do. Following this path will result in the big prize, far better services and much lower costs. If, subsequently, you choose to share these services you’ll also gain the much smaller saving from less of a common resource.
If you’ve already shared services, don’t worry, they can be redesigned wherever they are, and you will need the leaders to understand how to study and design; if they don’t lead it, it will never fly. A simple first step is to study how much failure demand there is, it will tell you how ineffective the services are and this should energise leaders to do something constructive about it.
If you’ve outsourced services to a private sector shared service then, probably, you’re already stuffed. Failure demand represents greater revenues for the private-sector provider as contracts are usually based on transaction volumes. Nevertheless insist that the provider joins you in studying the extent of failure demand and test their appetite for redesigning the services. Be open about the need to change the contract once a more effective design is developed. Going through the process will expose the attitude of the provider; are they a true ‘partner’? Be aware that getting out of the contract might cost a lot, as it has done for others.
When the shared services initiative started five years ago we were told it was going to save £2 to £4 million a year. We haven’t seen anything to match the promise; instead we are overwhelmed with evidence of failure. As time has gone on the promises become less ambitious. By contrast there is abundant evidence that public services can be massively improved while at the same time radically reducing costs and it can be done in months, not years.
While the shared services protagonists claim services will be designed around user needs – and often, would you believe, they use personas (imaginary people) and focus groups to do this – to design around user needs requires first of all thorough knowledge of citizen’s demands.
Shared services are IT-dominated. IT should be the last thing you do, not the first. Doing ‘Agile’ IT is merely doing the wrong thing faster. To separate technology management from operations management is a mistake. Technologists should be directly connected and subordinate to operations design. The techies like that, after all they think it’s cool to do stuff that works. You spend less and get more.
Demand is the big lever. Sharing services isn’t. You don’t need a business case or a plan. To improve services you have to study them – get knowledge – and then design them to be effective. Cast your mind over examples of ineffectiveness and ask yourself: if this had worked immediately for the user would it have been more efficient?
I encourage you to share this report with anyone who ought to care.
John Seddon
