IT Success Depends On Quality Of Spending, Not Quantity

If the hot buzzwords during the Internet era were things like "E-business" and "viral marketing," today's discussions are filled with phrases like "ROI" and "back to basics." That's because technology managers often need to prove that an IT investment will produce a return before they receive the money to spend on new systems.

But less than half of IT managers report a positive return on their IT spending, according to a survey of IT executives at 200 businesses in October and November, sponsored by Unisys Corp. The survey says about 44% report a positive return on IT investments, 42% report a level return, and 14% a negative return.

The survey concludes that the correlation between success and failure has little to do with the overall amount spent--and everything to do with how IT dollars are spent.

Josie Bradley, VP of IT for Mark Anthony Group of Vancouver, heartily agrees. "It still has to be a fair amount of money," she says, "but what you spend it on matters." For Mark Anthony, Canada's largest private alcoholic beverage company, strategic IT spending means moving away from trendy items such as E-business and back to servers and systems such as data warehousing. Bradley spent a lot, but won't say how much, on servers in 2001 to consolidate 19 servers to "four or five," with a goal of increased stability and uptime. She expects her IT spending to decline next year because of the investments made this year.

It's all a matter of being strategic with investments. The 14% who report losing money from IT investments may not be paying close enough attention to where the money goes, says Unisys VP Pete Samson. "There's often a gap between companies' stated IT priorities and what ends up being funded," he says. While IT executives don't always rate E-business as a high priority, it often receives a lot of funding. Conversely, increasing productivity and efficiency aren't always rated as highest priorities but consistently show strong returns when they're funded.