Strategies and plans are for the future. But if the future differs from the present, plans based on today’s world will become obsolete. One way to avoid this problem is to forecast conditions that will be in place when plans are actually implemented. Market and sales forecasts are typical for any unit, but technology forecasting is different.
The technology S-curve represents actual behaviour and measurements that can be made on the progres of a technology, usually of individual technical approaches to solving technical problem. Technology S-curves of performances generally assume an uppear limit based on physical capabilities determined by the underlying science of the area.
Diffusion of innovation can also be represented on an S-curve. The proportion of a population of potential adopters of a new technology can be plotted on the y-axis and time can be plotted on the x-axis of a graph. Here, the upper limit is the size of the population of potential adopters. For example, the proportion of the US merchant marine using mechanical power began to escalate rapidly after 1820 and slowed down dramatically after 1990, approaching 100 percent in 1960. The S-curve can be approximated with what is often called the growth curve of logistic curve, which is several alternative mathematical functional formulae.
Although the relationship between teh technology S-curve and technological forecasting with growth curves may be obvious, it may not be obvious why we would want to make the effort to forecast technology at all. We could ask the same question about sales forecasting. Although most companies do sales forecasting, many firms have more that one forecast in place. Marketing forecasts tend to be different from production or operations forecasts.
Operational plans often begin wit forecasts of demand required, quite independently of why demand is growing, in on a plateau, or is declining. Technology plans, then, ought to begin, at least in part, with forecasts about technology progress, quite independently of how or why these technological changes ought to occur. Dr.Martino says that anybody, any organization, or any nation that is affected by technology is, by default, entering into a forecasting exercise when resources are committed. By implication, the alocation of resources make assuptions about the technology future. The alternatives to forecasting systematically are all used periodically: no forecast (or future same as past); window blind (linear) forecasting; panic or crisis forecasting, or asking someone who has been successful in the past to forecast the the future again.
The virtues of using systematic technological forecasting is that these methods can be taught and mastered by people for cross-referencing, reviewed for soundness, and documented for learning when actual changes occur. If forecast are precise, they can be checked for accuracy, and even if they are incorrect, they can still be helpful, because a measure of forecasting performance is possible when the prediction is explicit.
There are foru basic methods of technological forecasting :
- extrapolation extension of a time-series pattern or trend, or incorporation of cycles is very useful for long term forecasting
- leading indicators act as barometers. Sometimes data are not directly available, so data on indicators like patens are often used
- causal models predict outcomes based on cause and effect. For example, scientist know an eclipse will occur based on the laws of physics.
- Probabilistic models produce a probability distribution for various outcomes.