Tools to Forecast Technology Innovations
Technology innovation can be forecast emotionally and empirically. Getting emotionally involved with a product or process can help in the initial growth stage, but then serve as a barrier when negative emotions (e.g., frustration) inevitably arise. Moving to data-based processes of forecasting prevent such emotional challenges. This article focuses on eight (non-TRIZ) forecasting methodologies. A companion article describes a case study in which TRIZ-based forecasting methods to successfully develop a product.
Making strategic decisions for product development is one of the most difficult challenges for the research and development (R&D) staff of a business. As Ralph Lenz, U.S. Air Force technology forecasting pioneer once said, “Technology forecasting may be defined as the prediction of the invention, characteristics, dimensions, or performance of a machine serving some useful purpose.”  How to decide between the optimization of existing technologies and the development of a new core technology? There is a high uncertainty related to these decisions and although many decision tools are available, and have been implemented successfully to various degrees, the decision-maker’s intuition is sometimes the only element for directing the company’s line of development.
Assessment of a company’s current technology should drive the direction of the R&D planning process. There needs to be a systematic process for assessing technology. [2, 5] There are eight (non-TRIZ) forecasting methods to aid assessment:
- Intuitive forecasts
- Consensus methods
- Delphi method
- Application of a statistical model
- Application of a causal model
- Analogy method
- Structural modeling
Intuitive forecasting is a very popular and widely applied methodology. “Asking the expert” usually provides the information for the forecast. It is assumed that the experts’ base of experience and education is sufficient, in a particular field, to predict or forecast the vectors of expansion or evaluation. Records indicate the pronounced fallibility of this forecasting method.
British author and inventor Arthur C. Clarke described some false predictions based on the “expert” intuitive forecasting by unquestioned authorities in his book, Profiles of the Future, including, for example, that in 1956 the Royal British Astronomical Society predicted that “space travel is utter bilge.” [1, 3, 4]
A simple method of overcoming some of the disadvantage of intuitive forecasting is the use of a “panel of experts.” The presumption is that many experts are more likely to be accurate than one.The U.S. Department of Defense has used this method successfully on projects including the Strategic Defense Initiative program (more commonly known as “Star Wars” as well as numerous weapon systems and vehicle platforms. The U.S. Air Forces Project Forecast was an effort in the mid-1980s to characterize high-leverage future military technologies and is an example of a large and successful application of the consensus method.Of course, the shortcomings of the intuitive method can be compounded during the application of the consensus method (the possibility of a hidden bias being canceled by a hidden counter-bias may not materialize).
To improve on the intuitive and consensus methods, the Rand Corporation developed the “Delphi Procedure” in which a panel of experts (like the consensus method) arrive at a consensus but eliminate the regulating of committee bias by employing a series of questionnaires. The first phase asks for the panel’s forecasts. The replies are compounded. The second phase requests comments on the Phase I compound forecast. Phase III is a derivative of Phase I based on the results of Phase II. A typical process includes five or six phases. The goal of this multi-phase process is a forecast convergence. Figure 1 shows the convergence of a hypothetical date for a certain even based on a multi-phase Delphi process.
Application of a Statistical Model
A statistical model is based on a series of observations of the phenomenon and it delineates the pattern of the association between the various factors (or variables) of the phenomenon that are of interest. Descriptive models used in forecasting are often quantitative, but qualitative ones are used as well. Many events, such as descriptive phenomenon, are single occasion events and as such they are a difficult phenomenon to model. Therefore, the application of a statistical model necessitates a thorough understanding – lengthening the forecasting process. (See Figure 2.)
Application of a Causal Model
A causal model is similar to a statistical model as it also describes (through research) the development of a phenomenon to be predicted. It is an improvement on the statistical model as it also provides the causative agent(s) for the occurrence of the phenomenon to be predicted. An understanding of the invariance of the phenomenon gives the forecaster good grounds for forecasting as this invariance is believed to remain valid into the future. A causal model is normally based on a population and is therefore valid only in that context. A causal model limitation is the necessity to assume uniformitarianism (invariance as a function of time).
A famous example of a large causal model was fabricated by the Club of Rome and published in the 1972 book, Limits to Growth. It consists of dozens of variables, including world population, birth rate, industrial and agricultural production, non-renewable resources and pollution. In the model, the levels, or physical quantities that can be measured directly, are indicated with rectangles, rates that influence those levels with valves and auxiliary variables that influence the rate equations with circles. Time delays are indicated by sections within rectangles. Real flows of people, goods, money, etc. are shown by solid arrows and causal relationships with broken arrows. Clouds represent sources or “sinks” (exits of material) that are not important to the model behavior.
The Club of Rome think tank started building their “World Model” by first constructing five sub-models. These concentrated on the five “basic quantities”: population, capital, food, non-renewable resources remaining (measured as remaining fractions of the 1900 reserves) and pollution. One of the sub-systems included the causal relations and feedback loops between population, capital, agriculture and pollution. Finally, the researchers combined all the five sub-models and created the final World Model, part of which is illustrated in Figure 3.
This method utilizes analogies between the phenomenon to be forecast and some historical event, or popular physical or biological process. To the extent that the analogy is valid (all analogies become invalid at a certain level), the initial event or process can be used to wake a prediction about future developments of a technology. (See Figure 5.) The technological forecaster uses the analogy method consciously and deliberately, examining the model situation and the situation to be forecast in considerable detail to determine the extent to which the analogy is valid. An example of this approach is delineated in the book The Railroads and the Space Program: An Exploration in Historical Analogy edited by Bruce Mazlish. The forecasters used 19th century railroad development as an analogous system to the U.S. space program. The utilization of growth curves is used to predict the advance of some technologies (analogous to biological or physical processes – the “‘S” curve, see Figure 4).
Trend extrapolation is one particular method of solving the prediction of the inflection problem associated with the “S” curve of the analogy method (when is the curve going to change slope, B). Instead of focusing on a single device and attempting to predict the future course of development of that device, the trend extrapolation method considers a series of successive devices while performing similar functions. These may be considered individual representations of a broad area of technology. (See Figure 6.) The extrapolated trend will eventually reach a physical limit and will lose its validity as the trend approaches this limit.
Structural modeling is an attempt to develop a mathematical or analytical model of a technology-generation process. As with mathematical models of any process, the purpose for model construction is to identify certain key elements, identify the functional aspects of those elements and express these functional aspects symbolically or mathematically. Structural models tend to be abstract and reductionist in their approach in removing what are denied to be non-essential functions. (See Figure 8.)
Each of the detailed eight methods is capable of forecasting technological innovation and furthering the overall innovation processes within a company.
- Clark, Arthur C., Profiles of the Future, Indigo, 1962.
- Gahide, Severine, Slocum, Dr. Michael S.and Clapp, Dr. Timothy G, “Application of TRIZ to Technology Forecasting – Case Study: Yarn Spinning Technology,”The TRIZ Journal, July 2000.
- Routio, Pentti, Arteology: Forecasting, 1996.
- Martino, Joseph P., An Introduction to Technological Forecasting, Gordon and Breach Publishers, 1969.
- Cowley, Michael and Domb, Ellen, Beyond Strategic Vision: Effective Corporate Action with Hoshin Planning,Butterworth-Heinemann, 1997.
Michael S. Slocum, Ph.D., is the principal and chief executive officer of The Inventioneering Company. Contact Michael S. Slocum at michael (at) inventioneeringco.com or visit http://www.inventioneeringco.com.
Catherine O. Lundberg is the director of operations of OnTech Operations Inc. She has seven years of experience utilizing customer requirements to drive new product development, manufacturing design, product enhancement, cost and waste reduction, and process improvement. Lundberg’s focus is on incorporating high level technologies and science with basic consumer interaction to produce products with exceptional customer acceptance. Contact Catherine O. Lundberg at catherine.lundberg (at) ontech.com.