New Maths: The Practice

Benchmarking yourself against previous investments and the investments of your peers is nothing new. But historically, the available data was hard to use for anything more than the highest-level comparisons. "Even then, the benchmarks were pretty worthless," says Merrill Lynch's Noble. And the data was insufficiently detailed to allow companies to compare specific aspects of their IT spending versus that of other companies. "There was a danger of comparing apples to oranges," Noble says.

But Noble believes that Rubin's two decades of benchmarking has created a database for the financial services sector that allows finer grained comparisons, a key step to understanding real trends over time and distinctions between competitors.

Rubin's own analysis of IT spending compared to a company's financial performance made him question the traditional measure of IT ROI, typically arrived at by dividing the IT spend into net revenue. The assumption behind that measurement was that net revenue should rise proportionally to the IT spend. Consequently, if revenue did not rise with a higher spend, the investment should be adjusted accordingly. Or more simply, the IT spend should be limited to a set percentage of net revenue. But such a calculation, in Rubin's view, ignores periods of investment and market change, which is why he suspects it hasn't correlated well with business results.

Rubin therefore began experimenting with other calculations and found that IT spending as a measure of business success correlated much better to another factor: operating expenses. He believes that comparing the IT spend to operating expenses better accounts for shifts in corporate realities: entering new markets, making previously deferred upgrades to the IT infrastructure and other capital improvements to competitiveness and efficiency, not to mention reacting to market shifts. While Rubin can't prove that this is the reason for the better correlation to business results that he derived from comparing IT spending to operating expenses rather than net revenue, the prima facie evidence comes from calculations across hundreds of companies in more than a dozen industries. So for companies looking for a yardstick to measure normalized IT spending trends over time, Rubin says his metric does a better job, especially because it employs two factors a business can control: operating expenses and IT spending. But enterprises still need to track the IT spend against net revenue, Rubin notes, even though revenue is, ultimately, not something anyone can control. "You don't want to spend money you don't have, so you have to be aware of that figure," he says.