Six rules for effective forecasting

[Latest News Headlines]

Look for the "S" curve and embrace things that don't fit. Be weak in strong opinions and know when to hold back. A look into Paul Saffo's latest thinking and the tools you'll need to build your "cone of uncertainty."

Be it the weather, the stock market, or next quarter's sales figures, the goal of any forecast is accuracy. Looking ahead, forecasters virtually always strive for numbers that are precisely on target.
But futurist Paul Saffo, writing in the Harvard Business Review, urges business forecasters to focus on "effective" rather than "accurate" forecasts.1
Forecasting, of course, is a discipline of likelihoods and probabilities. No one knows for sure what the future will bring. So the notion of "accurate" forecasting, Saffo argues, is a bit of an illusion.
The primary goal of forecasting then, is to "identify the full range of possibilities, not a limited set of illusory certainties."
"Identify the full range of possibilities, not a limited set of illusory certainties."
– Paul Saffo

Saffo observes that decision making in business ultimately relies upon intuition and judgment, and effective forecasting provides the "essential context that informs your intuition."
To make sure the context that the forecast provides is as helpful to your intuition as possible, and to avoid the trap of "illusory certainties," Saffo offers six "rules for effective forecasting."
Rule 1: Define a "cone of uncertainty"

The cone of uncertainty represents a range of possibilities that "extend out from a particular moment or event." Ideally, the process of creating this cone "broadens your understanding by revealing overlooked possibilities and exposing unexamined assumptions regarding hoped-for outcomes."
Rule 2: Look for the "S" curve

Forecasting demoDriver-based forecasts
Look for an S-shaped pattern (flat-steep-flat) of change across time. New technology, for example, is typically adopted slowly at first, then gains rapidly in popularity, and then levels off.
An illustration is the growth of television. It was first introduced commercially in the 1930s. Yet it took 20 years of fairly flat growth before it reached the "inflection point" of very steep growth in the 1950s.
Rule 3: Embrace the things that don't fit

Pay attention to indicators that don't fit into "familiar boxes." They may be the early signs of an "industry-disrupting S curve just starting to gain momentum."
Digital photography, for example, didn't fit anywhere in the traditional film and film-developing industry. Yet digital photography revolutionized—and nearly killed—that industry.
Rule 4: Hold strong opinions weakly

Don't ignore conflicting evidence and be ready to change course quickly if your strongly held belief is disproved. The remedy for an over-reliance on strong opinions is frequent forecasts that include new data.
Rule 5: Look back twice as far as you look foward

Look for historic parallels, and look well beyond the recent past. As Mark Twain said, "History doesn't repeat itself, but it does rhyme." Current trends may speed up or slow down, but they will rarely continue on unchanged.
Rule 6: Know when not to make a forecast

"History doesn't repeat itself, but it does rhyme."
– Mark Twain

When market-changing or world-changing events occur, it may be best to withhold speculation on what will happen next. When the Internet burst on the scene in the late 1990s (after decades in the academic world) many forecasters said that the old economic rules no longer applied.
As it turned out, the death of the old economy was, to quote Mr. Twain again, "greatly exaggerated."
Good rules + good tools = effective forecasts

Some of the above rules for effective forecasting are best suited to the realm of macroeconomic and geopolitical events. But others apply equally well in the familiar context of daily business. And that is where performance management solutions come in.
Use analysis to chart trends, and planning to build "what-if" scenarios.

Using analysis, for example, forecasters can chart the progress of industry trends and define the "cone of uncertainty" for whatever scenario needs to be considered.
Powerful planning tools for "what-if" scenarios can help forecasters distinguish the possible from the probable, the outliers from the likely.
And Cognos solutions are specifically designed to enable best practices such as rolling forecasts, to make sure that today's strong opinions don't obscure the indicators of tomorrow's altered business environment.
The upshot? You spend less time on the mechanics of forecasting, and more applying your intuition and judgment to the universe of likely outcomes.

1Paul Saffo, "Six Rules for Effective Forecasting," Harvard Business Review, July/August 2007

More...