The note stated that

While government agencies and departments are looking at shared services as a way to reduce costs and increase efficiency, some of the potentially shareable services will be supported by cloud-based solutions

and that

While not suggesting that government organizations that are currently engaged in devising a shared-service agreement among themselves are doomed, it is important to understand where the cloud will impact government shared services and where it will not and when.

providing advice about which types of shared services were likely to be challenged (or “obscured”) by cloud-based services, which ones would have a chance to provide sustainable value, and which ones would be an intermediate step toward more centralized services (where the difference between shared and centralized is that in the former clients do have a say by participating in the governance process, while in the latter they are just clients).

This being said, it does not look like governments have been taking much notice. There are quite a few shared service initiatives (or strategies) that focus on IT infrastructure services, which that are presumably seen as a commodity by most prospective clients. Examples include the Canadian initiative to consolidate a whole-of-government infrastructure and run it as a centralized service, or the more recent US federal “Shared First” strategy (a Gartner research note is being prepared), which is looking at commodity IT services as easy wins for shared services, despite some of them being already targeted through an earlier “Cloud First” initiative.

On the other hand, according to NextGov, the US Congress is directing the Department of Defense to use commercial cloud computing services rather than those provided by the Defense Information Systems Agency (DISA).

As stated in the congressional record dated December 12, 2011:

Not later than April 1, 2012, the Chief Information Officer of the Department shall submit to the congressional defense committees a performance plan for a reduction in the resources required for data centers and information systems technologies Department-wide [including] Migration of Defense data and government-provided services from Department owned and operated data centers to cloud computing services generally available within the private sector that provide a better capability at a lower cost with the same or greater degree of security

I assume this will be subject to negotiation and interpretation, but exemplifies quite well the potential – or, better, the actual – conflict between cloud computing and shared services.

Anybody who is familiar with the complexity of government IT (and very much so in the defense environment) knows quite well that migrating to any cloud solution, and even more so when commercially provided, requires much more than turning a key and pulling a plug. But the real question here is whether shared services at an infrastructure, commodiity level have a sufficiently long shelf life to allow the expected returns on the investment to materialize: in fact, also getting the shared service levels and governance mechanisms right is far from trivial.

The race between shared services and cloud will be interesting to watch, as shared service providers move up the value chain (from infrastructure to applications) and commercial cloud-based offerings mature along that same value chain. What is important is not to be obsessed with either, but to create fair conditions for departments and agencies to procure the service they need at the best value for money from the most appropriate service provider.

Today many governments around the world are either mandating or pursuing or planning the establishment of shared services as a way to reduce operational costs. Be it infrastructure, applications or entire business processes, it is quite easy to build a business case for why services should be common inside a particular jurisdiction, rather than duplicated in each and every agency or department.

On the other hand, it appears that shared services are having a hard time. Queensland, Western Australia, South Australia, Victoria have all had their fair share of issues with shared services, and this is happening quite consistently in other parts of the world. More recently, many have been looking at the bold consolidation decision by the Canadian federal government, which has set a rather aggressive timeframe to consolidate over 300 data centers into 20 and over 100 email systems into one.

There is no doubt that duplication of services and spending across different government agencies makes little sense. But experience shows that there is insufficient attention to the governance aspects (agencies want to have a say in how the shared services are structured) and – more recently – to technology evolution that is making some of the technology that is being targeted through shared services more and more commoditized.

The challenge is no longer to put one organization in charge of delivering shared services, but to look at how to support agencies in efficiently buying the same service from external service providers. In this respect, I am impressed with how New Zealand is moving to support agencies in buying infrastructure-as-a-service, by selecting preferred vendors that agencies can buy from. This is not dissimilar from how GSA in the US and the UK Cabinet Office are doing, but with an important difference: while in the UK and in US there is a data center consolidation initiative funning in parallel to the cloud one (and agencies may choose either cloud or the consolidated data center), New Zealand seems to be going for cloud only, still leaving agencies free to choose within the boundaries of a framework contract

This balance between control and choice is at the basis of the success (or lack thereof) of future consolidation and shared service initiatives.

The topic of both conversations was how to evolve their mission and strategic objectives recognizing that all government departments will be struggling with tough budget cuts for the foreseeable future.

While in this particular jurisdiction e-government has been quite high on the agenda, also to close the gap with other jurisdictions in the region, there has been a sudden awakening to the reality of much tighter budgets and closer scrutiny.

Whether these central organizations realize this or not, there is a shift of priorities from electronic government to affordable government, from providing better services to being able to maintain existing service levels, from debating enterprise architectures to making sure resources are available to pay salaries.

Of course investing in technology and greater automation helps reduce costs in the longer term. But resources for those investments are not available, and the focus must shift from improving processes to radically changing the way outcomes can be achieved.

This requires a very deep reflection by business owners about the role of technology, how to harness the power of individual workers enables by inexpensive, consumer technology, and how to squeeze all possible value from their IT departments. It also requires those IT departments, as well as any central or shared service organization, reflect about the value it adds to individual IT departments and its business users.

Some of these central government organizations have the feeling that they will become more important going forward. After all, centralization prevents duplications and enforces standardization, doesn’t it? However the pace at which technology is evolving toward commodity solutions casts doubt about the value added by centralization at the lower levels of the technology stack (e.g.. infrastructure). Further, the effort required to set effective governance structure around shared services, and the resistance met to centralize functions and assets, are unlikely to realize cost savings in the short term, especially if compared to the economies of scale that some vendors can realize on very large client bases.

Whole of government CIOs and central or shared services organizations may actually become less rather than more relevant if they don’t chose the right battles to fight and win. And those battles may be very different from those they set themselves for fighting. In most cases, they have to figure out how to exercise more power with less control: not sure many are ready for this.

This is far from surprising of course. The economic recovery is between uncertain and fragile, government budgets keep being under scrutiny, greater public spending to further ignite recovery will further affect spending on government service delivery and operations. As I have described in research note From Modernization to Survival: Radical Cost Reduction in Government (available to Gartner clients only – a press release with some of the content is available here), government organizations are at a point where they need to think more creatively and radically to how to save costs in a sustainable way.

At the same time, there is a need to innovate to respond to evolving citizen and business needs. As people face challenges like globalization, long-life learning, shift from permanent to temporary work arrangements, immigration and emigration, governments need to transform their services and the way they work.

Therefore the emerging problem is how to balance innovation, which usually costs money, with cost reduction.

Now, what is the best way of striking that balance across one or multiple jurisdictions? Should this being achieved through greater consolidation or by empowering individual agencies?

We decided to take this as the topic of a point-counterpoint session that will be held at our upcoming US Symposium in Orlando. My colleague Jerry Mechling and I will debate it during the public sector track that complements the official Symposium agenda, with sessions taking place over breakfast and lunch.

Here is the description of our lunch session:

Balancing Cost Containment and Innovation Today: Does It  Take More or Less Centralization?

In tough economic times government organizations are called to reduce their costs while maintaining their service levels. This implies that they have to innovate service delivery models and business processes with very limited discretionary spending. The session will discuss and debate two different approaches: one based on greater centralization and consolidation of IT services to reduce recurring costs and free resources for  innovation, and the other one based on empowering individual departments and agencies to strike that balance, limiting centralization to where it is absolutely unavoidable. The audience will be asked to actively participate in the debate, where each of the analyst will support and defend one of the approaches

It would be great to collect some viewpoints from this blog’s readers about whether more or less consolidation would help find the balance.

In one case, the CIO is overseeing a significant consolidation and centralization program that will rationalize computing assets and applications from several different agencies . While a shared services organization has already been formed, he will need grater power to mandate the use of certain services, through the issuance of new regulations in the coming months.

I was positively surprised, as I have met in the past government CIOs who exude the power given to them by political mandates and directives, as opposed to those who have to make their way through consensus building. This particular CIO seems to be very well aware that a political mandate or a directive won’t necessarily make him too many friends, and that those who may look supportive of his increasing authority may turn their back after the necessary regulations are passed.

When I suggested that he should focus on the 20 percent of client agencies as well as services that he wants to really win, he quietly smiled and gave me the impression that he had already thought through that and had his clear list of priorities in mind.

The second CIO operates in a heavily federated environment, where agencies still have a great degree of autonomy in how they source IT. We had a great conversation about how to prioritize e-services, and how to look at citizen participation as a way to improve operational efficiency more than a way to deliver better services.

As he looks at the future of his central organization and the opportunities and challenges posed by new sourcing models, such as cloud computing, he seems to be smartly playing to become an excellent procurement and advisor organization, rather than a service delivery one.

Both CIOs had this in common: a relentless focus on how their organizations can add value over time. Sustainability of the services they provide is for them as important as the immediate business case that leads to more centralization as a reaction to difficult economic times.

Interestingly enough, both CIOs come from recent private sector experience in technology companies and have not been in the job for long. People with a similar profile, used to environments where decision making is considerably faster and authority can be exercised with less horse-trading, often lack the articulation to navigate the difficult waters of government organization. This is not the case for either of them, and I hope that they will be able to keep a steady course of action in the coming months and years.

This has been mostly driven by the need for greater efficiencies. In fact, having different infrastructures, applications and processes across different departments and agencies makes little sense when what those are trying to accomplish are pretty much the same things.

The move toward centralization has been accelerated by technology trends such as virtualization and cloud computing, which pave the way to better utilization and rationalization of otherwise distinct resources and push a “commoditization” agenda for enterprises in all industries.

In the government domain, several jurisdictions are moving toward some form of “government cloud” activity, although in different shapes and forms. National governments in the US, UK, Singapore, Israel are pursuing the adoption of cloud computing on a large scale, while countries like Denmark or states like Victoria or Queensland in Australia are on the path for consolidation and centralization of IT infrastructure.

Is it all gold that glitters? Is centralization going to deliver all the value it promises? And, if it does, at what cost?

Responding to a recent post by my colleague Michael Maoz and in my post about the Australian spin of the Gershon review I have casted doubt about centralization being all good and fair.

Let me be clear. I am not saying that centralization of IT assets and processes in government is a bad thing, on the contrary. What I mean is that we need to understand very carefully its risks as well as its benefits. In my post on Gershon I have already mentioned its possible downsides, and here I would like to focus on two in particular.

The first one is constraining sourcing strategies for agencies and departments. Once they have a single, centralized service provider they are in a monopoly situation: either by mandate or by applying budget pressures, centralized IT services agencies are made the only service provider that agencies can chose. On the other hand, it is quite possible that they may find more cost-effective and flexible service offerings from external service providers, which either specialize on service organizations of their size and nature or just can offer lower prices for the sheer size of their infrastructure. A centralized service provider needs to accommodate the requirements of entities (i.e. government agencies) that are very diverse in size, mission, nature, and it is very unlikely it will be able to provide the most cost-effective services to all of them. Indeed a smart centralized service provider will be very good at sourcing its own services to the market, but with increased commoditization of services it will be challenging to prove its value added with respect to individual agencies going straight to those external service providers.

The second point, which I appreciate is probably even more controversial, is about the role that government inefficiencies play in making IT markets work. Especially during an economic downturn, and as a consequence of the relentless search for efficiencies by the private sector, government becomes a very important market for many IT players in all jurisdictions. As I have stressed in my research about the geopolitics of cloud computing,  in spite of globalization and commoditization, governments will inevitably prefer to spend in their own jurisdiction. The more they centralize, the less likely they are to rely on local, second or third tier IT service providers. The more they centralize, the fewer their buying centers, the fewer opportunities for the IT industry to serve the government market.

For those who say that this line of reasoning is very close to subsidizing an industry, I’d like to recall recent examples in areas like the car industry, where government funding of all sorts (lending, consumer incentives, etc) have kept the industry breathing.

So my point against centralization is that it must be taken with a grain of salt. More autonomy in sourcing and purchasing to individual agencies does not necessarily mean more waste but may – in some cases – make government more efficient as well as respond to transient and longer term needs for economic development.

Sir Peter Gershon is the person who led efficiency reviews in the UK in 2004 and the in Australia, in 2008, triggering a wave of cost-containment and rationalization measures that are deeply changing the way governments deal with IT. In the UK, the Operational Efficiency Program in 2009 stemmed from the early work he did, bringing his recommendations to the next level. In Australia, on the basis of the lessons learned in the UK, he proposed much more aggressive cost saving targets that have really set things in motion.

Therefore one would argue that what he has done is good, which I would certainly agree with. However, amongst all these praises, I would suggest that his recommendations – and the way they have been implemented – have also created some major drawbacks.

The first one is that as management focuses almost entirely on cost saving and rationalization, it leaves little room for innovation, or puts it on the backburner.

The second one is that it supports the assumption that “centralized is better” and that “government can be run like an enterprise”. In this respect, I am particularly concerned with some of the developments that are taking place both at the federal level and in some of the states. For instance the state of Victoria, and – to a lesser extent – Queensland, are pursuing a much stronger centralization strategy than they did in the past. In the former case, a specific organization (CeniTex) has been created , which operates almost as an independent service provider, with a board of governors without any direct representation from departments (which sit in a stakeholder advisory committee, hence with a much less loud voice). Interestingly enough, the “dogma” of centralization may soon be challenged by the reality of commoditization, when individual departments may access technology services from cloud service providers which may be priced more competitively than anything a central provider can offer. Unless departments retain some discretionary power when it comes to procurement (which runs contrary to centralization) they may not be able to get those additional savings.

The third one concerns how IT investments are decided. Australia has a long tradition of effective investment frameworks, both at the federal and at the state level, to drive balanced decisions that take into account different dimensions of public value, i.e. efficiency, effectiveness, citizen service, mission alignment, and so forth. It is not by chance that in a report (access to Gartner clients only) where I looked at good examples of “public value of IT” frameworks around the world, two out of seven came from Australia. The focus on efficiency will inevitably skew decision towards efficiency as being the primary if not the only driver, hence leading to a potential imbalance in investment portfolios.

Last but not least, one of the highlights of Gershon reviews has been that saving money on government operations frees resources to be invested elsewhere. As in Australia investments in broadband will exceed Gershon savings by a factor of 20 to 30, and nobody really knows how beneficial those will be, this casts doubt about how “efficiency” and “effectiveness” ought to be measured.

The UK government has to be praised for relentlessly looking at how to improve its administration’s efficiency, through the Gershon review first (which has been somewhat imitated by Australia) and the Operational Efficiency Program now. The report recognizes and build on the results of actions taken in response to the Gershon review.

Gartner is working on a research note for its clients that examines in detail the recommendations from the report. Two areas I have been looking at already are back-office and IT and local incentives and empowerment.

The former calls for building savings achieved after Gershon in department budgets, for embracing benchmarking disciplines, for strengthening governance of IT, for furthering and consolidating shared services, for developing internal IT capabilities and profession. The latter calls, amongst other things, for accelerating joint working in local areas, for improving service design by leveraging frontline and other government workers’ expertise, for embedding innovation and collaboration in capability reviews.

These are all important areas, and some will require strong drive and continued effort to be implemented. What seems to be entirely missing in the report, though, is any reference to socialization (i.e. how to involve constituents in service delivery and redesign as well as in operational improvement) and to commoditization (i.e. how to leverage non-government-specific infrastructures and services to further reduce cost and free management resources to better focus on mission-critical tasks and processes).

Why isn’t there any link to what the Power of Information Report says about using public data to engage people in helping government improve its performances? Why isn’t there any reference to something like cloud computing, IT utility or other forms of IT commoditization?

Last but not least, the report advocates greater consolidation and centralization of IT decision-making. While this makes sense in general terms, how  are contributions of individual departments to mission-specific innovations going to be encouraged?

It will be necessary to work on the overlaps, complementarities and conflicts among this report, the Power of Information Report and the Reform Agenda (see my previous post as well).

Jurisdictions worldwide have published  government interoperability frameworks in multiple versions (see a few examples from the UK, Germany, Australia, Denmark, New Zealand) and all these have helped move technical interoperability in the right direction. Compatible technical architectures or document standards are necessary conditions for processes and applications to interoperate, but they are not sufficient. As the European Commission’s work on the European Interoperability Framework shows, there are several level of interoperability, and technical is just one: the toughest ones are semantic (do data have the same meaning?), organizational (are processes compatible?), legal (do similar laws apply to the same issues?).

This reminds me of when I was doing research on software reusability, back in the late 80′s. At the time, barriers to software reuse were technical (i.e. different platforms, operating systems, programming languages) as well as organizational (i.e. reuse cost models, design methods and processes). However even with most of those technical challenges overcome (by the adoption of open standards and more and more technically interoperable platforms), reuse is still relatively low, because organizational challenges are far more difficult to solve.

The same applies to interoperability: we can have all countries and regions using the same (open) standards, and yet data models may be incompatible. Technical interoperability provides a common lexicon and part of a common syntax, but can’t help with semantics. The realm of semantic and organizational interoperability dangerously borders whole-of-government enterprise architecture, another venture that has rarely provided much value, besides creating a context for compliance and scrutiny.

An interesting question to pose is: how much more effort will be put on cracking the toughest interoperability problems, taking into account the difficult times ahead?  This question would clearly be amenable to some scenaric planning, but let me just take the most controversial view. If this is going to be a long and deep recession and governments need to deploy exceptional resources to sustain economic development and social cohesion within their respective boundaries, government interoperability may become irrelevant.

Let me start with the counterargument to my own thesis. In order to rapidly achieve challenging political objectives around economic recovery and job protection and creation, different government agencies and tiers need to collaborate more effectively and efficiently. Focus on national and local priorities will cause efforts like “European Interoperability” to slow down or even grind to a halt. But also within countries, states, provinces, what is likely to happen is a recognition that there is no more time to devote to semantic or organizational interoperability, to government-wide enterprise architectures and global transformational programs.

On the one hand governments will have to slash the cost of their operations (finance management, HR, procurement, general administration) and put some programs on the afterburner (as they contribute less to economic recovery or welfare). On the other hand they will have to focus on reinforcing critical programs and launching new ones (such as bail-outs for entire industrial sectors or new infrastructure investments) of a nature and a scale that are almost unprecedented.

This will lead to consolidating and commoditizing IT related to general administration , wherever feasible, by imposing shared or centralized services: interoperability won’t help much (besides using legacy data), as agencies will need to transition toward common services, applications, infrastructures, as opposed to running their own in an interoperable fashion. Where this is unfeasible or requires too much time, things will be left as they are, just reducing budgets.

Low-priority programs will be left lagging behind, with just the IT support that is required to let them survive, but little or no money for any significant enhancement, let alone major architectural redesign.

Finally, high-priority programs will have very specific interoperability requirements that are likely to break the boundaries of current efforts in the field. They will have to engage resources, information, processes in different industry sectors. For example, how to measure the impact of several dozens or hundreds million dollars? How to prevent misuse of such funds? How to adapt job creation and support measures to the changing landscapes in various sectors and parts of the jurisdiction?

Interoperability will become (or – in a certain sense – remain) a tactical issue, to be faced in conjuction with a specific problem to be solved. There will not be much space left for one-size-fits-all measures, any more enterprise architecture, or religious battles about which flavor of an open standard is open enough.

Interoperability will still play a role to help cushion government agencies from the risk of defaulting IT suppliers responsible for maintaining legacy products. In many cases, these products will be proprietary in nature and hardly interoperable with new products that are candidate for replacement. Therefore, backward compatibility with proprietary data formats will become a key criteria in selecting new, open-standard-compliant alternatives.

In essence, in a long and deep recession the attitude to interoperability will become much more pragmatic. This is an opportunity to show what its true value is, and a challenge for those who still pursue too an ambitious and (sometimes) abstract approach to it.

[...] 5.3.3 Consider shared services carefully

The review received a number of inputs from industry indicating the significant benefit to be obtained from shared services. In respect of back office applications, however, I also considered the mixed experiences reported to us by the CIOs of a number of states and territories, together with the recent experience of the UK Government as described in the May 2008 UK National Audit Office report Shared services in the Department for Transport and its agencies. In light of this, I have concluded that moves towards back office shared services between agencies should only be undertaken on a very carefully selected and controlled basis. In the meantime, as a first step towards a wider adoption of these arrangements, I consider it essential that all agencies quantify both the back office service levels and the associated costs of their current provision arrangements, and that they use this as the basis for determining what improvements can be realised through their own efforts, such as process simplification and a reduction in manual interventions. This will help create a stronger foundation on which to assess the additional benefits that can be obtained from moving to a shared service in the future. [...]

 

The bottom line is that agencies should benchmark what they are currently doing, and improve themselves, before deciding whether it makes sense to enter into a shared service agreement.

But while they do so, they may find out that externally sourcing some of their processes could be a good fit, irrespective of whether there are other agencies to share those services with. In other cases, they may discover that collaborating with other agencies ( which ranges from buying together, to sharing processes and assets, to centralizing them) is the way to go.

The economic downturn will call for grater sharing and consolidation, but it is important to dispel the myth that shared services are the solution to most problems and recognize that – sometimes - they make problems even more complicated.