Ensuring Better Business Forecasting: Forecasts are always educated guesses, but too many companies set off on a vain quest for precision. Better results come from embracing and trying to quantify uncertainty.

• Author: Meirav, Udi
• Position: Forecasting

Although significant earnings shortfalls now seem routine, initially, they blindsided some of the best and brightest CFOs -- along with just about everyone else. These across-the-board failures indicate a widespread breakdown in the reliability of business forecasting. Executives and analysts may bemoan the lack of "visibility" as a new, unfamiliar challenge, but in fact, the fog has always been as thick. It's just that the road was once somewhat straighter.

To understand the failure of forecasting, we need to reexamine the way businesses think about the future. Forecasting is famously difficult and inaccurate, but ironically, one reason for that inaccuracy is a misplaced emphasis on precision. The first step toward more insightful -- and more accurate -- business forecasts is a frank acknowledgment that the future is neither fully predictable nor an entirely blank slate. Further, we must recognize that forecasts are not meant to give us a sense of certainty, but to guide us to the best decisions in an inherently uncertain environment. To obtain better forecasting, financial executives can do three things:

* Understand the link between unreliable forecasting and suppression of uncertainty.

* Implement a systematic embrace of uncertainty in all planning and communications.

* Lead the adoption of forecasting practices that consistently measure uncertainty.

Consider the volatile semiconductor business in the 1990s. For the makers of DRAM memory chips back then, the living was easy and the margins were high. A chip fabrication plant -- or "fab" -- was practically a license to print money, and at a rate of hundreds of millions of dollars yearly. Manufacturers were rushing to build new fabs, at close to $2 billion each. Rosy forecasts abounded, and the industry had been so highly and consistently profitable that claims were being made about the end of the infamous 'Semiconductor Cycle.'

Texas Instruments (TI) -- a leading DRAM manufacturer at the time -- was concerned about investing in a new fab, despite the optimistic forecasts. Heated boardroom debates took place over the risk of future price changes. That was nothing unusual, except that TI's executives realized that uncertainty is not merely an inconvenience or an opportunity, but also a quantitative and valuable form of knowledge. That recognition led to a historic shift for one of America's most successful semiconductor companies.

Let's look at how forecasts deal with uncertainty. Despite...