None of my Computerworld columns has generated as many comments as the Sept. 6 analysis of the computer paradox. Readers were troubled by the absence of any correlation between IT spending and corporate profits, which is perhaps the most accurate way of defining and measuring the paradox.

The letters noted that rapidly falling prices of desktop computers, as well as the explosion in Internet use, were self-evident proofs of productivity gains from computerization. The fact that none of these developments has so far shown up favorably in corporate financial data didn't seem to discourage anyone.

The bottom-ranked 45 firms showed better results than the top 45

A leading IT magazine has come up with a novel response to the computer paradox. It has suggested that IT-based innovation is the source for creating new business value. The magazine ranked 500 U.S. corporations according to their innovative capabilities. Corporations engaging in e-business, enterprise resource planning, data mining, enterprise portals or similar buzzword-compliant activities received gold, silver or bronze "medals." Four gold medals put a company in the top rankings.

I could argue over the process by which individual firms were awarded their ranks, because the findings reflected the unverifiable opinions of the IT people who were surveyed.

But I won't, because we need not depend only on surveys to assess business value. Audited financial reports offer more reliable indicators. When IT executives spend money, they are acting on behalf of the firm's shareholders. Therefore, the most appropriate metric for judging business value from computer-based innovations would relate them to net profits, operating profits and return on shareholders' equity.

To verify the trustworthiness of the magazine's innovation rankings, I devised a way to compare top- and bottom-ranked corporations. The question was whether a company awarded four gold medals would deliver greater business value than one with four bronze medals.

Because only 450 of the magazine-ranked firms published their complete financial results, I chose the top 45 ranked firms (such as Marshall Industries, Microsoft and Sprint) and compared their average financial performance with the bottom 45 (such as American Greetings, Tidewater and Textron).

How well did the innovation rankings compare with the financial measures of value creation? They didn't hold up:

• 1998 operating income growth over 1997: The bottom 45 outgrew the top 45, 18.3% to 15.4%.

• Five-year average net income growth: The bottom 45, 14.8% to 14.3% for the top.

• Five-year average net equity growth: Again, the bottom 45 outgrew the top 45, 18.5% to 16.2%.

• Five-year average employee growth: The top 45 finished ahead, 10.2% to 8.5%.

So, the bottom-ranked 45 firms showed better results than the top 45 in three of the categories that measure financial performance. There were other interesting indicators, but I included only the five-year employment growth comparisons. When it came to hiring, the top "innovators" exceeded the growth rates of the bottom-ranked firms. The innovators grew faster in personnel but not in business value.

It's often argued that the benefits of innovations show up only in the long run. That may be true about many start-ups, but the rankings discussed here cover only well-established firms whose shareholder values depend on historical as well as current financial performance.

Where I've served as CIO, I've always favored innovation. But the current frenzy to adopt new computerized practices without examining the cash benefits can be damaging, especially if a leading publication blesses such a point of view. As in any arms race, the speed and expense for adopting computerized solutions must be tempered by sober economic analysis. For a business to survive, it must be steadily profitable and increase shareholder value at rates superior to those of its competitors. Magazine surveys that rely on surveying opinions about technological excellence aren't credible if the financial results don't support such claims.