This analysis of the Western Australian Government’s shared services woes is by Steve Hodgkinson, the director of analyst firm Ovum’s government practice in Australia and New Zealand. Prior to joining Ovum he was the Deputy CIO for the Victorian State Government in Melbourne, where he was responsible for e-government and IT strategy across the State Government’s departments and agencies. He led a four-year programme of activity to establish the Office of the CIO and implement shared services and infrastructure consolidation initiatives.
opinion The decision by the West Australian (WA) government to abandon its shared corporate services is a salutary reminder of the governance realities of the Westminster system of government. Portfolio and agency autonomy is the dominant force, whatever the desires of central agencies and the grand plans cooked up for them by consultants. Just because benefits appear compelling in a spreadsheet does not mean that they can be realised in practice.
The key lesson from the Queensland, WA and South Australian shared services misadventures is that it is foolish to underestimate the depth and resilience of agency autonomy. Agencies have strong incentives and drivers — supported by culture and on-the-ground operational realities — to make decisions based on local imperatives. Whole-of-government incentives and drivers, on the other hand are weak. When push comes to shove, agency autonomy always wins because it is agency heads that carry accountability for operational service failures.
Unfortunately, when shared service are forced onto agencies this autonomy manifests itself as an inability to compromise requirements for shared systems leading to complexity and risks. Theoretical economies of scale rely on one-size-fits-all systems and fail when the resulting customised system is so compromised that it fails under its own weight.
Governments are well advised to treat consultants’ shared services business cases with extreme caution. Shared services can work, but the risks are high that costs will be greater, and benefits smaller than expected. Success requires realistic expectations and strong leadership.
The way to keep risks to a manageable level is to constrain the scope of shared services to commodity-like infrastructure services. Victoria’s CenITex shared services agency is a good example. No agency can mount a credible argument that it has unique requirements for data centre ops, application hosting and core desktop services. Risks increase when the shared service attempts to provide business systems that need to be tailored to distinct agency requirements. Expecting agencies to agree to participate in a forced march towards common business requirements is a triumph of hope over experience.
As cloud computing services mature in Australia, governments would be well advised to take a good hard look at the cloud for shared infrastructure and applications. How can it be that over 100,000 organisations around the world meet their needs using Salesforce.com — a one-size-fits-all shared service — while a handful of agencies in WA cannot agree on how to share common HR and finance systems after wasting millions of dollars trying?
The key difference with the cloud is “cloudy is as cloudy does” — the services and apps already exist; agencies can make their own decisions. Economies of scale come as a service attribute and are not contingent on all agencies being herded into one service by the Department of Treasury and Finance.
IT strategists and procurement executives often express reservations about the risks of cloud computing while ignoring the even greater risks of flawed shared services strategies. It is time for them to wake up to the cloud.