I first detected the existence of explosive IT growth predictions last fall. A senior researcher from Gartner announced that the growth rate of IT expenditures will accelerate, reaching - on average - 10% of U.S. corporate revenues by 2005. I don't usually pay attention to percentage-of-revenue benchmarks because they vary from 0.1% for natural resources firms to a New York investment bank that spends 10% of revenue on IT (spending more than $70,000 on IT per employee). The idea that U.S. averages would approach by 2005 what I consider close to a maximum spending limit was so far-fetched that the prophecy couldn't be dismissed.

The 10% prediction didn't go away. Within a few weeks, magazines quoted it as the new standard for assessing how well a company keeps up with competitors. A leading weekly then announced on its cover that U.S. firms could expect IT spending to rise, on average, to 8% of revenues by 2004 as a prerequisite for survival. After that, the can-you-top-this forecasting competition was in full bloom. An article in a popular bimonthly magazine announced that by 2004, average spending could be as high as 25% of revenues. Finally, a respected computer magazine featured a consultant's assertions that average spending could be expected to reach 20% to 40% of revenues by 2004!

One way to test the credibility of any forecast is to give it a sanity check. If true, does it make sense? I examined spending statistics for 8,961 U.S.-listed corporations with total revenues of more than $5 trillion. I then projected estimated average IT spending based on 8%, 25% and 20% to 40% of revenues.

The results were astounding. Here's a table of what we would have to spend just to match the predictions:

Projected 2004 IT/ Revenue Ratio	Median IT Budget/ Employee	Median IT Budget/ Total Compensation	Median IT Budget $ Millions - 2004	Avg. Growth Rate
8.0%	$15,865	33%	$10.5	24%
25%	$49,578	104%	$32.8	118%
20 to 40%	$69,410	145%	$45.9	174%

None of the three IT-revenue ratio forecasts is believable. Their projected average IT spending growth rates of 24%, 118% and 174% would surpass even the most optimistic projections in employee payroll as well as any other corporate indicator, including profits, head count and administrative expenses. Further, the future IT spending numbers are in current dollars, and IT costs are supposed to be decreasing, according to Moore's Law. Therefore, it defies comprehension to try to understand how the amount of actual computing power we would be buying would increase profits sufficiently to pay for the enormous IT bills.

When next year's budget plans come up for review later this year, it may be foolish to refer to spending forecasts to justify accelerated spending. By all means, spend money on IT if there's a demonstrable payoff. But beware of those who inflate bubbles beyond realistic limits. Somebody who's reviewing your spending plans may puncture that bubble.