AdSlot 1 (Leaderboard)

IT budgets – How to slice the pig

Paul Brock, MD of IT at New York-based investment bank Bear Stearns, implemented a new system in the bank last year. Nothing novel about that, you might say. But the system, acquired from Ejasent of Mountain View, California, wasn’t meant to help the bank carry out its business. In fact, it had nothing to do with banking, finance or even administration. The system, explains Brock, was intended to ensure that the bank’s individual departments were charged for the amount of computer storage their activities consumed.

In the world of IT, it’s called chargeback – charging back the cost of IT to the users who drive those costs. Bear Stearns developed the chargeback model in order to manage what its IT expenditure actually delivers. In the process, it has put Brock – and other IT executives like him – on the frontline of a tense tussle between user departments and IT functions over who pays for what.

Basically, chargeback is a form of absorption costing. As such, it’s prey to the chief failing of any absorption scheme – deciding the basis of the driver that determines the extent of the absorption, cost by cost.

“The simplest way of doing it is by headcount, where the total cost is divided by the number of users and then split by the number of users in each function,” says Brock. The problem, he adds, is that this assumes users make even use of IT resources, which is not the case.

Function heads whose users consume vast amounts of IT resource don’t complain, of course. But department heads whose users consume significantly less IT resource than average soon object to the hit to their budget.

“Users quickly push you towards a usage-based model, which is fine in theory, but more complex to administer,” says Brock.

Brock explains that “usage-based chargeback forces people to confront their usage and decide if it is optimal”. And based on the bank’s experiences with chargeback, he’s expecting a 25%-30% reduction in storage costs as users shift their behaviour.

More popular in the US than in the UK, chargeback polarises businesses and their IT and finance arms. Some, like Brock, see it as a useful way to optimise resource allocation. Others see it as a drain on resources, where the money spent administering the system could be spent delivering value to end users.

“If you consider a small business, its IT costs are reflected in its profitability,” says Simon Bragg, European research director at Bedfordshire-based ARC Advisory Group. “It has to pay for the IT resources it uses. Why should a large business be any different?”

With IT expense retained within an overall IT budget rather than charged back to departmental cost centres, companies can’t get a handle on the real marginal cost of winning business, says Bragg. “To boost their department’s profitability, there’s nothing stopping a function consuming as much IT as it can – even if the cost is higher than the revenue it produces.” Chargeback, he says, forces users to reconcile within their own budget the marginal cost and profitability of new business.

That’s one argument. But there’s an opposing view, chiefly centred around the practicality of calculating a sensible chargeback figure. “The theoretical merits of chargeback are obvious, but the fact that people are still talking about it as if it were a new solution is testimony to how difficult it is to administer,” says Paul Clermont, former management consultant at Nolan Norton.

At the moment, the wrangling about chargeback has a distinctly US flavour.

Its use is most appropriate for organisations with a large IT infrastructure, making it more appealing to giant US organisations. But the debate is moving to Europe, partly because companies want to get a better handle on what their IT expenditure actually delivers. The e-commerce and ERP boom of the late 1990s was driven by often overblown competitive pressures and issues such as Y2K. It all seemed necessary and affordable. The decision to adopt a back-office system such as SAP or PeopleSoft would be taken at a corporate level, not a departmental one. But opinions have shifted and chargeback is starting to look like an idea whose time has come.

If so, then companies contemplating its use need to find a way of doing it sensibly. Rakesh Kumar, VP of analyst firm META Group in Fleet, Hampshire, argues that an ill-applied chargeback scheme is counter-productive and can actually drive up costs. Take, for example, a server running both a finance application and an HR application. Each application’s needs in terms of uptime, reliability and so forth are radically different.

Dividing the cost by two and charging it to the respective functions merely disguises the problem. “If it had been outsourced to a company such as EDS or IBM Global Services, they would have differentiated between the two availability requirements,” says Kumar.

One way round this is to focus on what users actually need. Bragg recommends the use of service level agreements, basing the chargeback scheme on these.

The trouble is, taking a purely IT-based look at business activities can produce set cost drivers that don’t build to a coherent whole that can be divided up and charged back. One department’s chief concern might be uptime, another’s might be processing speed. Nor do these drivers necessarily equate to costs. A function might require limited uptime and communications bandwidth, yet have huge storage requirements. A headcount-based solution, or one based on mainframe MIPS (millions of instructions per second), is at least based on a common metric and produces some sort of solution, albeit not necessarily ideal.

One company that has found a way around this is Manchester-based Co-operative Insurance Services (CIS). Historically, CIS’s approach to chargeback was fairly rough and ready, explains project accountant Andrew Booth. Carried out on a spreadsheet, it boiled down to charging back on the basis of mainframe MIPS consumed and any clearly identifiable software development costs. Wanting to achieve both greater accuracy and pin down the IT costs associated with certain products, CIS decided to exploit the activity-based costing capability of a system from ALG Software of Knutsford, Cheshire. The plan, explains Booth, was to piggyback on to an existing timesheet-based system that was filled in by everyone in the IT department. Mainframe MIPS would still be used as a proxy for the business’s hardware requirements, but the timesheets would provide an accurate basis for charging out every single hour of IT staff’s working day. And to provide visibility at product level, the only thing needed would be to cross-link the project codes used by IT staff with the products that those projects were destined to support.

With each department now seeing a monthly IT charge in their management accounts, the result has been a sea change in attitudes toward IT. “People are now asking questions about their IT costs. We’re also seeing fewer IT projects initiated because people are aware they’ll be charged for the work and are asking if it’s worth it,” says Booth.

In the US, such sentiments are driving some chargeback enthusiasts to construct IT procurement contracts that permit IT consumption to vary as demand changes. At Merrill Lynch Global Private Clients in Princeton, New Jersey, the bank’s director of business programmes and solutions, Marvin Balliet, explains that having been on the receiving end of the bank’s former chargeback scheme as CFO of its 8,000-strong technology division, he was given the task of coming up with something better. The result was a scheme that employs software from San Francisco-based Business Engine, which combines labour costs with non-labour cost drivers that business people can understand and control. The result is much higher elasticity in the consumption of IT services.

Consequently, the bank’s approach to buying IT resources is changing.

“We need as much flexibility as possible,” says Balliet. For example, if the bank comes up against a capacity limit, extra capacity can be leased on a three-month basis to determine whether the additional demand is permanent or transient.

For chargeback critics who carp that weighing the pig doesn’t make it any fatter, Merrill Lynch’s experience certainly provides a sobering counterpoint.

Related reading