Here’s an alarming statistic that I read recently: 81 percent of everyone surveyed thinks their IS organization is average or below average. If “below average” translates to “below the mean,” only 20 percent of us are in the top 50 percent.

Since human perception is a pretty dull scalpel, “average or below average” may not be quite as precisely defined as “worse than or equal to exactly half the total number.” Let’s try a different interpretation. Figure anything within one standard deviation of the mean counts as average. In round numbers about two-thirds of any sample falls inside one standard deviation. The remaining third splits in half, so one sixth of any sample is above average. The remainder – five sixths, or just over 83 percent, are average or below.

Mystery solved! The 81 percent who figure their IS departments are average or worse are almost exactly the number who ought to think so according to the inviolable laws of statistical sampling.

The authors of the paper reporting this statistic made no further comment, so we don’t know if its absurdity escaped them or not. That three college professors who specialize in business metrics resorted to this kind of number, though, speaks poorly of the state of the art in IS measurement. I certainly didn’t do anything like that in my new book, Bob Lewis’s IS Survival Guide from MacMillan Computer Publishing (nor would I ever stoop to shamelessly plugging it in this column).

As we found last week, we have plenty of measures to choose from, all internal ones that tell us how good our processes are compared to internal baselines or external benchmarks. What we lack are external measures that assess the value we create for the enterprise. The measures we have tell us, to borrow a phrase, whether we’re doing things right, but not whether we’re doing the right things.

We do have one external measure at our disposal. The cost of technology is depressingly easy to measure, and our detractors gleefully proclaim it during budget season. But the value we create? That’s a lot tougher.

The purpose of any measurement system is improvement (ignoring its important use in political self-defense). The point of calculating the value we deliver to the enterprise is helping us increase it. How do we create useful measures of value? It’s tough. At the highest level, the formula for calculating value is Bang per Buck. We know how to measure the buck part, which leaves the bang as the part we need to measure. Start by listing the major categories of benefit we provide:

• Capabilities needed so the company can achieve its strategic goals.

• Capabilities needed for effective marketing efforts.

• Fully automated (and therefore high-efficiency) processes.

• Capabilities needed by redesigned processes.

• Capabilities for improving communications with customers and suppliers.

• Capabilities for improving internal communications.

• Capabilities that allow individual employees to be more effective in their jobs.

See a trend? Except for the rare situation that allows for complete process automation, the value we deliver is capabilities. They’re enablers – necessary but not sufficient conditions for success. To measure the value we deliver, we need to understand how to measure the value of a capability when that capability may or may not be used effectively.

How will we go about that? In principle, we need to list every contributor to success in each of these categories, then assign a weighting factor to each of them that reflects its relative importance or contribution.

Great theory. Can we turn it into practice?

Oh, gee, we’re about out of space. Too bad … you’ll have to tune in next week to read the next installment.