Forecasting

forecasting

can you see into the future?
how can we reason about the unknown?

on this page

> unreliability
> benefits of forecasting
> what to forecast
> steps
> techniques
> links

on other pages

> estimation

home page contact us

“If we do not take the time to review the past we shall not have sufficient insight to understand the present or command the future: for the past never leaves us, and the future is already here.”

Lewis Mumford

Purpose benefits of forecasting	Form what to forecast
Process steps	Material techniques

Unreliability of forecasting

It all depends why you're forecasting in the first place. If you start with a fantasy of eliminating any risk associated with the future, forget it.

According to Charles Smithson, the financial engineering guru, "In markets as efficient as the financial markets, attempts to outpredict the market are unlikely to be successful."

He goes on to argue that "Because forecasting is thus a very unreliable means of eliminating risk, a remaining alternative is to manage the risks."

We agree with Smithson about the importance of risk management. But that doesn't mean you shouldn't even try to forecast.

Benefits of forecasting

Forecasting is valuable for two reasons.

The end-result itself can sometimes be valuable, because an accurate forecast may improve the quality of the plan(s) based on it. It may help identify and evaluate risks, clarifying factors and reveal assumptions.

More often it is the forecasting process that is valuable, because it sensitizes managers to change and to the relevant factors, helps them understand what is likely to occur, and builds a consensus.

Sometimes the consensus is crucial. The fact that everybody is agreed on a forecast may be more important than its accuracy. Particularly if the consensus involves many organizations or units, crossing organizational and cultural barriers.

Consensus works both ways. An agreed forecast can be self-fulfilling. But beware the self-disfulfilling prophecies!

What to forecast

Some variables may be easier to forecast than others - for example, aircraft miles flown depends not only on passenger miles flown but also on size of aircraft, thus if you don't know trends of aircraft size, you are likely to get a better estimate of passenger miles than of aircraft miles.

Sometimes it is easier to forecast a ratio than an absolute figure.

Steps of forecasting

There are four steps to forecasting:

1	Set objectives for forecast - what decisions will forecast influence, and how - what aspects of plan are vulnerable to surprise - how accurate must forecast be to be of any use?
2	Obtain forecast - from existing sources or commissioned study.
3	Evaluate forecast, including assessing the sensitivity of planned actions on the accuracy of the forecast - what exactly is being forecast - what assumptions does the forecast itself make - source of forecast and past reliability - what objectives of forecasters, and how do they compare with your own?
4	Disseminate the forecast, persuade others of its accuracy and appropriateness.

Techniques of forecasting

Techniques can be divided into ‘exploratory’, which project forward from the past/present situation, and ‘normative’, which trace backward from a hypothetical future situation to assess its likelihood, timing and consequences.

	extrapolation	Extend or bend a known line or curve from the past into the future. This requires some theory to decide whether to extend or bend (e.g. straight line, S-curve). Don't forget to consider trends, seasonal & cyclical variations.
	precursor	To estimate the success of a new product, consider what happened last time we introduced this kind of product, and adust the estimate for any plausible differences. Or if historical figures are available from other firms or periods, (especially the rate of technical substitution). Use with caution.
	leading indicators	Find internal or external variable(s) that is shown to be correlated to the variable for which a forecast is desired, except that the correlated variable either moves sooner, or can be measured sooner, than the desired variable.
	model building	Develop symbolic representation or simulation for prediction, to forecast several interrelated variables together (thus avoiding the dangers mentioned under extrapolation, above).
	scenarios	Use alternative pictures of the future, not as fixed predictions but as frameworks for discussing future possibilities (use brainstorming, synectics, lateral thinking, ...).
	delphi	Attempt consensus among experts by iteration of individual forecasts.
	judgement	Choose subjectively which forecast(s) to believe.

Links

veryard projects > forecasting > links

Some excellent material from the Whole Earth Review archives.

	Foresight
	Future	Donella Meadows, Spring 1999
	Scenarios

top

home page