Earlier this year, I did a blog entry looking at the problems of forecasting the future. Forecasting, I said, was not so much a matter of predicting the future. It was more about exploring patterns of interaction that allow people to see things differently.

In the latest edition of the Harvard Business Review, Silicon Valley forecaster Paul Saffo elaborates on this in his interesting piece Six Rules for Effective Forecasting .

Saffo's six rules are:

1. Define a Cone of Uncertainty : Map out all the factors, including the relationships between them. Distinguish between the highly improbable and wildly impossible, include wild cards and don't make the mistake of drawing the cone too narrowly. Doing that will get you blindsided.

2. Look for the S curve: Most change follows an S curve pattern. Change starts slowly, and incrementally moves along quietly, then suddenly explodes, and eventually tapers off and drops back down. The trick is to define the S curve before the inflection starts. And don't expect the opportunities to be the same as those everyone has predicted. In the early 80s for instance, PC makers predicted that every home would have a PC on which people would do word processing, and use spreadsheets and read encyclopaedias on CD. But when home PCs became common, most people were using them for entertainment and when they read encyclopaedias, they were going online.

3. Embrace the things that don't fit: Keep in mind indicators that don't fit into boxes. Because we hate uncertainty, we dismiss them as failures when they don't get up straight away. Saffo points out that the earliest graphical antecedent of Second Life was Habitat which disappeared quickly. But it came back in another form with the rise of multiplayer online games.

4. Hold strong opinions weakly : Don't rely too much on seemingly strong information just because it supports your conclusion. Saffo says good forecasters forecast often, and are always trying to prove themselves wrong.

5. Look back twice as far as you look forward: The recent past is rarely a good indicator of the future. But what you need to do, he said, is look for the patterns. It's been said that history doesn't repeat itself but sometimes it rhymes. The forecaster's job is to find the rhymes, not identical events. And it's important not to dismiss past events that don't fit in with your view of the present and future. One of the best examples is Iraq. The planners at the Pentagon had assumed Iraq II would unfold like Iraq I. They dismissed Vietnam because the US had lost that war. The result: deja vu.

6. Know when not to make a forecast: There are moments when forecasting is impossible. The cone of uncertainty can contract and expand and sometimes it becomes so wide, that anything can happen. That's when the wise forecaster will refrain from making a forecast at all.

You can read Saffo's piece here


no comment untill now

Add your comment now