Weill maintains that the traditional model of having only a thin layer of shared services -- with every department otherwise having their own IT processes -- is untenable because it drives up legacy maintenance costs. Under CISR's alternative model, all publicly available infrastructure and most shared and standard applications should be shared services. Data management processes unique to a department or business unit should remain separate.
Large or long-standing organisations may struggle with this step, due to the huge number of extraneous IT systems and processes that organisations naturally tend to accumulate. However, Weill says executives should be forever working to weed out the extraneous processes. "I believe that [streamlining] is within our power and would argue that if we don't do it, it is malpractice." Using Weill's model helps to sort the wheat from the chaff by identifying the extraneous processes.
Weill admits that shared services can be a tough sell to division or area heads. "Anybody who's worked in any shared service in an organisation will know as soon as you approach someone who runs a division and say, 'We'd like you to follow this shared service way of doing things', they say one thing to you -- 'Oh no, we're different'." But Weill says if you can convince division heads that costs will go down and efficiency will increase, even the most reticent among them will soon see the benefits.
CISR's model also has benefits for innovation. Weill recommends innovation occurs at the division-specific level. "When innovation occurs on top of the platform, it's got a high failure rate," he says. But this means innovations that do succeed are effective and worth incorporating into the organisation as a whole.
"[Innovative processes] that fail should be cleaned up. The ones that succeed then drop into the infrastructure if they have enterprise-wide capabilities. This gravity process is the way of dynamic learning in an organisation -- innovate, succeed, drop into the infrastructure, consolidate, free up money, innovate again."
Having effective accountability mechanisms in place is vital for the success of this model. "If you don't have accountability to figure out whether or not there are pay-offs, then you never know what was good about the innovation and what wasn't."
DECISION 4: How good do your IT services need to be?
"You get what you pay for but we can't fund everything," Weill says. "One of the things that we decide in IT is how good our services should be."
This question is best answered not by CEOs but by heads of individual divisions. "Most organisations, government agencies and others have services that have SLAs," Weill says. "Division heads can then pick and choose the best level of service they can achieve within their budgets, emphasising the services that the division uses the most."
Join the CIO Australia group on LinkedIn. The group is open to CIOs, IT Directors, COOs, CTOs and senior IT managers.