22
Technology Forecasting

Short Description

Technology forecasting aims to provide information about the direction and rate of technology changes. It uses logical processes to generate explicit information to help industry and government anticipate practical, ecological, political, and social consequences of developments in technology. In government, this information is used to inform policy. In industry, the information can be used to inform strategic improvements to, or replacements of, products and processes and predict changes in markets.

There are four elements in a technology forecast:¹

• A time horizon (either the time of the forecast or the estimation of time when the forecast should be realized)
• A specific technology
• Some parameters to the technology (characteristics and capabilities gauging level of performance)
• A probability statement about the outcome or range of outcomes predicted

Technology forecasting is performed using a variety of techniques.

Background

Technology is defined in the Oxford Dictionary as "the application of scientific knowledge for practical purposes." A forecast predicts or estimates a future event or trend.

For millennia humans have been fascinated by the future and what it might hold. The interest of business in systematically attempting to predict the future is more recent.

There are two broad perspectives that may be taken on what it is that prompts the development of technology. One is that technology will develop in response to scientific and technical opportunity (a technology is discovered, and an application for it is then sought), the other is that technology will develop in response to the need or desire for change (an application is found, and technology is developed for that purpose).

As scientific knowledge changes, so does technology. Change in technology has implications for many aspects of modern business. While some technology change amounts to refinement of existing technology, other change can render existing technology obsolete and can have immediate effects on your ability to compete.

The development of technology experienced an explosion in growth after the Second World War. The defense and space industries were subject to heavy government investment, and research undertaken in these led to many developments that had significant impacts on other industries. For example, the work in the space industry on miniaturizing electronics revolutionized the production of domestic appliances.

Following on the heels of the radical growth in technology development came a major upheaval in the nature of commercial competition. For many decades, economies around the world grew at a fairly steady rate. Change tended to be gradual and the marketplace fairly predictable. However, since the 1970s, markets became much more volatile and unpredictable. Competition became fiercer, and changes in marketplace dynamics occurred much faster than in the past.

Together, these changes, the rate of change itself, and the intensifying of competition in the market, have made looking to the future in terms of technology crucially important to the competitiveness of business and of national economies that is to both the public and private sectors. A change in technology may prompt modification of existing government policy or development of new policy, both of which may have flow on effects for industry; for example, by making tax concessions available for specific areas of research and development or by supporting the setting up of a new industry. Other considerations for industry include that a change in technology could mean a sudden loss of market share as products are superseded or to the loss of a market where a technology is completely replaced—for example, floppy disks and floppy disk readers.

The modern practice of technology forecasting as used in business has, as previously mentioned, its roots in the U.S. space and defense industries in the 1940s and 1950s. It was used by the U.S. as a tool to keep its technology ahead of the Russians during the Cold War.

There exist a variety of methods used to forecast changes in technology. The oldest is expert opinion, and while it is not all that widely used today, refinements of the method survive (for example, the Delphi method) and are discussed later in the chapter.

Probably the earliest systematic technique for forecasting technology change to find its way into business is morphological analysis. The process was developed by an astrophysicist, Fritz Zwicky, in the 1940s as a way to systematically invent solutions to specific problems. The process was first adopted and applied to future studies for use as a corporate learning tool in the 1980s.

In the late 1950s, another important technology forecasting method was developed. The Delphi method originated with and was refined by the RAND Corporation to enhance its ability to conduct business in the defense industry throughout the 1950s and 1960s. It uses the consensus opinion of a panel of experts to explore technology advances.

Over time, technology forecasting techniques have been adopted more generally by business as a tool that makes available information directly relevant to managing a firm's investments in technology. Concepts related to technology forecasting include technology road-mapping and foresight.

"Technology road-mapping" is a term invented by Motorola to describe its method for developing technical strategy. It is a process used by businesses to plan for the projected needs of the marketplace. It provides a plan of action for organizing research and development activities over a course of years (usually no more than 10 years) in order to achieve the stated goal. By focusing on a goal in the future, road-mapping helps a firm to allocate its investment resources and technology capabilities and focus its activities on strategically achieving its goal.

Foresight studies are usually undertaken by a national government to identify and encourage the development of desirable technologies. Foresight studies may play a role in developing the national economy by luring international research and industry firms to set up for business in particular nations.

Strategic Rationale and Implications

As the technology used by business has become more complex and business has become more technology reliant, many firms are actively monitoring their technology requirements and the technologies they rely on to try to stay ahead of the game (or at least keep up).

Technology change potentially has implications for all business. The products a firm markets, the processes a firm uses for production, or the equipment it uses to provide its service may be superseded, thus giving a competitor an immediate advantage in the marketplace.

Technology forecasting can provide information with obvious and immediate applicability and with significant cost implications. For example, the products or services a firm sells are vulnerable to the effects of technology change. Considerable investment is involved in developing a new product and setting up production lines. This will be lost if the product is rendered obsolete by technology developments. It is also possible to gain a price advantage over competitors by investing in more efficient technology production processes.

Figure 22.1 is an example of a traditional business model when compared to the potential impact of an alternative technology.

Figure 22.1 Online distribution versus CD technology in the recording industry
Source: Adapted from "Boosting the Payoff from R&D," by R.N. Foster (1982), Research Management, 15(1), pp. 22–27.

Additionally, the internal functioning of a business relies on technology. The equipment and software used to conduct administration and distribution processes and the communication methods used in most firms to reach clients are all based on particular technologies. Often this technology will be introduced at great expense and will commit the firm to a considerable substitution cost. For example, consider a situation where a firm installs a new software system to record and track customer service calls. Besides the expense of the software and associated licenses, the firm will have to invest in training for staff. There may also be expenses incurred in migrating earlier records to the new system. A major change or update to a new system will incur expenses all over again.

Historically, there was a lag of around five to seven years (generalizing across industries) between the making of a discovery and the embodiment of the invention into a practical application. There is then a further time lag before the innovation will have an impact on the market.² Recently, these times have tended to shorten.

Technology forecasting aims to extend the time during which a firm may work on its response to a new technology by giving the firm insight into change before it happens or before it has a practical impact. This buys time for the research and development of new products and/or services. It will also give some insight into the likely longevity of investment returns a firm can expect from its investment in a particular technology.

There are a variety of methods used to generate technology forecasts:

• Expert opinion may be sought about likely directions for change. Delphi is an example of a process using expert opinion to forecast likely change. It explores future technology developments by drawing a consensus opinion from a panel of experts.
• Trend extrapolation and growth curves use information from the past to predict developments likely in the future.
• Morphological analysis uses information about current technology to try to find new applications for existing inventions.
• Relevance trees systematically break down a problem as a method for finding a solution.
• Monitoring follows current research and finds links between inventions to predict what practical innovations may arise from them. An example of monitoring is patent analysis.³
• Historical analogy picks an analogous technology from the past and plots development in a new technology as following a similar growth trajectory. Historiographical analysis is discussed in detail in Chapter 25, "Historiographical Analysis."
• Scenarios can explore future technology by presenting a series of perspectives on possible futures each involving slightly differing conditions to arise.⁴

Technology forecasting predicts future developments by anticipating the probable characteristics and timing of technology. It will focus on one specific technology outcome and explore the likely attributes of that technology at some nominated time in the future—for example, in 10 years' time. The most useful technology forecast will include some sort of estimate of how the likelihood of its predictions materializing. It should always make explicit the assumptions on which it is based.

While technology forecasting itself is not necessarily concerned with a firm's profits, it should provide sufficient insight to allow a firm to make informed decisions about its investments in technology, which will in turn have a direct bearing on future profitability.

To not undertake any form of technology forecasting is to assume that either technology change is not relevant to a firm or that the technology used is static.

Strengths and Advantages

The greatest strength of technology forecasting is its ability to inform current and future investment decisions throughout a business. Undertaken early in a project, it can provide valuable information about the likely longevity of a technology. It may even indicate the time period over which a widespread (even market standard) technology is probably going to be replaced.

Technology forecasting is a flexible process and can be tailored to investigate the precise area or areas of possible change directly relevant to an individual business or industry. Its usefulness is not limited to highly technical industries and its results are readily comprehensible.

The variety of methods available offers a range of sophistication and allows a firm to choose a method appropriate to its budget. It can be integrated into a firm's regular planning processes or conducted as a one-off process for a particular project.

The individual methods used for technology forecasting all have their own particular strengths.

The Delphi method allows a firm to tap into the expertise of experts across a range of specialized fields while protecting the resulting forecast from the subjective biases or "blind spots" of each individual. The consensus opinion from the panel of experts gives each expert the opportunity to advise on a situation and then to revisit and refine their advice in light of the opinion of their peers.

Trend extrapolation can use available statistical data to assist in the development of indicators and/or inferences about the future rate of change. While this method is often used to predict change in one aspect of a technology, it can be used to pull together information about a variety of aspects and predict a plausible and possible future direction.

Growth curves can also usefully predict when a technology has reached maturity and is likely to be replaced by something new.

Relevance trees are a powerful stimulus for thought. They provide a systematic process for finding a solution to a problem. A relevance tree would identify relationships between parts of a technology or process and its potential development.

Morphological analysis gives detailed analysis of the current and future structure of an industry and shows existing and potential gaps. It leads to explicit consideration of solutions to fill gaps in the market. The process, while exhaustive, is precise.

Monitoring of patents and general research trends can give a firm advance warning of likely new inventions that may be significant for its business.

As can be seen, technology forecasting is a flexible process, providing a range of methods that can be adapted to suit a firm's budget, resources, and timeframe.

Weaknesses and Limitations

As with any form of forecast, the usefulness of a technology forecast is heavily dependent on the quality of the information and the validity of the assumptions upon which it is based. The aspects of technology to be considered in a forecast must be carefully chosen so that important factors are not accidentally missed. The ability to understand what are the important factors driving a particular technology's development may require expertise in that technology that is beyond a firm's own staff. It may be the case that there are shortcomings built into the process as a result of the culture of the firm itself.

The reputation of the individual championing the project or outside consultants presenting the information may sway a firm's willingness to accept and use information and cloud its ability to interpret the complexities of a forecast. For example, a forecast confirming the firm's own preconceptions and emanating from a highly reputable source may be given greater weight in decision making, which is out of proportion with the parameters of the forecast.

Technology forecasts do not provide hard conclusive results. The forecasts will give a prediction of the probable attributes or appearance of a technology at some inexact time in the future.

Technology forecasts involve exercises in predicting probability, which is a notoriously difficult to do. There are several common errors we make when attempting to judge the probability of an event occurring. For example, a memorable event may seem more likely to recur, even though it may be memorable for being unusual in the first place. Generally people tend to overestimate low-probability events (for example, have you ever bought a lottery ticket?) and underestimate high-probability ones (for example, ignoring the very likely negative consequences of one's own pet vices). Human beings are also highly likely to allow their own personal experience or anecdotal evidence to distort their perceptions of reality. In fact, acting on a technology forecast ultimately requires a leap of faith.

Additionally, the individual analytical methods used for compiling a technology forecast all have their own weaknesses.

There are several points at which the Delphi method may fail. The appointment of properly qualified experts is crucial to the integrity of the opinion that results. If the panel or a portion of the panel is not experienced in the area you are investigating, the opinion will not truly be an expert one. Similarly you must be careful that the questions you are asking the experts will answer the specific questions you want answered about the future.

The Delphi method is structured in a way to minimize the impact of idiosyncratic responses; however, there is no way to control the amount of time and care taken by any expert taking part. It is possible that the time-consuming revisiting of questions during the Delphi process may even prompt less time and less care be taken with each round. Whether or not this is the case, the quality of the result depends on the quality of the responses and the range of knowledge of the experts used.

For trend extrapolation and the use of growth curves, the biggest weaknesses are the assumptions underpinning them—that is, the future will follow the patterns of the past. There is also the implicit assumption that change is not sudden.

The processes of trend extrapolation and growth curve plotting are profoundly dependent on the limits chosen for the analysis. Donnelly notes an example where limiting analysis to a particular technology when extrapolating a trend and ignoring other information from the marketplace resulted in disastrously inaccurate predictions. He cites the example of television manufacturers during the 1950s and 1960s. The overall trend was for television sets manufactured in America to become larger and more like a piece of furniture, which led the American firms to concentrate product development in this area. However, at the same time, the Japanese manufacturers were starting to make (and consumers were purchasing) compact sets. The real trend in the market was for greater variety in the size of TV sets, and the American firms effectively locked themselves out of a significant portion of the expanding market by relying on simple trend extrapolation.

Typically, trend extrapolation will look at the future from the perspective of one factor of change at a time. This assumes minimal interaction between different technologies and different technological developments. In fact, it is often the case that change is driven by interaction between aspects of technology.

Relevance trees and morphological analysis are subject to human error and vulnerable to lack of insight on the part of those constructing them. Both can be very time consuming to construct.

Morphological analysis does not take into account factors external to the particular problem or technology in question—for example, costs. It also requires knowledge of all possible solutions to a problem in order to find new applications for the technology (to solve the specific problem you have). Without knowing all alternatives, the analysis is compromised. As all possibilities must be represented, time must be spent in listing many impossible alternative uses for technologies.

Relevance trees are a very general approach for solving a specific problem. It can lead to pursuit of a fundamentally flawed course of analysis, as the flaws may not be obvious until very late in the process.

Each technology forecasting technique including monitoring requires diligence to be effective. Incomplete monitoring may well end up being misleading about likely developments. Similarly, simply monitoring developments, but not analyzing what is found, is worthless.

Process for Applying the Technique

The first step in applying any technique for technology forecasting is to identify as best you can what it is you and your firm's decision makers wish to predict or look for. Are you looking to explore the future for technology change that is driven by pressures of competition and opportunities arising from current technology research? This may be a driver if your firm is seeking to improve its current products to keep ahead of competitors. Or your questions may be more goal-oriented so that you are looking for technology development as a response to some need you have; for example, to fill a gap in your product range or that your clients have. Are you planning to make one technology forecast as a background for a current project? Are you expecting to put a regular process of generating technology forecasts in place?

Whatever the drivers, you and your firm's decision makers need to be clear up front as to the parameters of the analysis. If not, you will be caught in a mire of information, facts, and opinions that will ultimately lead to biases, blind spots, and ineffective analytical outcomes.

Technology forecasting is performed using a variety of techniques, and five common techniques are addressed next.

1. Delphi Technique—Expert Opinion

The Delphi technique builds a technology forecast based on expert opinion. It uses a consensus of opinion to try to minimize the effect of individual bias.

The Delphi process uses a panel of experts, chosen for their knowledge of a particular field or issue in question. If the questions being asked are general, then the panel should have representatives from a variety of disciplines. For example, if your objective is to understand the potential take up of a new technology by society in general, you might approach experts with not only technical backgrounds and industry experience, but those with interests in social areas such as design, cooking, and gardening, to name a few. You may also need to involve a large number on the panel, say in excess of 20.

Where you have specific questions to put to a panel, you are most likely to want experts with specific experience—for example, specialists in the particular technology and possibly experts from outside but relevant areas. Your panel could involve around 10 to 15 experts.

A facilitator coordinates the process and sends a questionnaire or survey to each of the experts on the panel, often in the form of a series of hypotheses about when and which scenarios are likely to occur and seeking responses to them. Often the expert is asked to respond to a scenario by answering a series of questions using a Likert scale (where, for example, circling the number "1" indicates strong disagreement with a statement, "2" indicates disagreement, "3" indicates a neutral response, "4" indicates agreement, and "5" indicates strong agreement). While it is possible to undertake a Delphi process in a face-to-face setting, it is essential to the process that the experts be allowed to respond anonymously. Anonymity is important to prevent pressure being placed on participants to respond in any particular way.

Once the expert opinions are all received, the responses are collated by the facilitator. The results are then sent back to the experts, showing them statistics on points of agreement and conflicting opinion (anonymously) and seeking a further response. Sometimes written arguments may be submitted (anonymously) with a detailed opinion of why some judgment is right or why it is misguided. The experts are invited to respond again to the survey or questionnaire in light of the statistical feedback.

The aim is to find consensus, meaning a majority agreement. Experts who find that their response to a particular question is out of step with majority opinion may choose to revisit it if it was an opinion they felt uncertain of the last time around.

This back and forth procedure will continue for a given number of rounds (usually three, as studies suggest that this gives the best balance between achieving reasonable consensus and not exhausting the goodwill of the panel) in an attempt to build a consensus of expert opinion. The opinion is usually presented at the end of the process as a proportion of experts agreeing to a particular response—say 80% of respondents agreed strongly that such-and-such change is likely to occur in the next five years.

However, the experts are not to be pressured to find consensus if it would compromise their considered opinion. If consensus cannot be reached, final distribution of responses will appear in the forecast with a note that it does not represent a consensus of opinion.

2. Trend Extrapolation

Trend extrapolation requires a forecaster to consider change over a period of time, understand the factors that have driven that change, and predict future change from this knowledge. It is used to forecast change in functional capabilities. This method relies on an assumption that past drivers of change will continue to influence the future and ignores short-term fluctuations in trends as it aims for a long-term forecast. It is useful in an environment where development tends to occur fairly constantly.

Generally, statistics (numeric data about past developments) are plotted onto a graph against time. A line is roughly fitted to the points plotted. It may be straight or curved (for example, showing an exponential growth). The mathematical formula that best explains the shape of the line is then used to predict the position of future points on the line (roughly); that is, over future points in time. See Figure 22.2 as an example of trend extrapolation.

Figure 22.2 Trend extrapolation
Source: Adapted from Lawrence, S.R. (2002). "Technology Scanning & Forecasting," University of Colorado, http://leeds-faculty.colorado.edu/lawrence/mbat6450/docs/schedule.htm.

Limit analysis may then be used to check the utility of a trend extrapolation plot. Limit analysis is based on the fact that all technologies have a limit at some point beyond which there can be no further improvement. For example, improvements to the braking system on a motor vehicle will stop a car more quickly; however, nothing can stop the car instantly (that is, without the elapsing of any time at all). The mathematical plot of the improvement will continue the line (as a theory) beyond the point where any practical improvement is possible. Extrapolation will not yield any useful information if applied to a technology already close to the limit of its potential.

Trend extrapolation may be used to forecast future developments in a technology that has a precursor technology (or several precursors) with a known path of change. The shape of the curve for the precursor (or precursors) is used as a guide for the shape of the technology in question. For example, plotting the efficiency gains in Formula-one car racing engines would give a shape that correlates to efficiency improvements in domestic motor vehicle engines. This process enables more complex predictions to be made.

Trend extrapolation may also examine past developments and predict future ones on the basis of judgment, rather than using statistics and graphs. This will give less precise results than a graph; however, the results may nevertheless be accurate as a forecast. This method is particularly suited to situations where numeric data is too complex to plot into lines, such as where many different factors contribute to the issue being forecast.

3. Growth Curves

The growth of development in technology change is thought to follow an s-curve, similar to the growth of biological life.⁵ This shape is regarded as universal and is used extensively in plotting product life cycles (see Figure 22.3).

Figure 22.3 Generic s-curve

The s-curve illustrates the gradual process of research leading to a new invention, which is then improved upon (where the line goes most steeply upwards) until the limit of the technology is approached (and the line levels out). See also Figure 22.4.

Figure 22.4 S-curve model
Source: Adapted from Foster, R.N (1982). "Boosting the Payoff from R&D," Research Management, 15(1), pp. 22–27.

The stages of Figure 22.4 are

1. Embryonic—The first few years of R&D yield low returns as the focus is on a wide range of research and knowledge acquisition.
2. Growth—Critical Knowledge starts being applied and developed, causing the productivity of R&D to skyrocket.
3. Maturity—The productivity of R&D begins to wane as the technology reaches its natural limit.

Plotting the development of a technology over time as an s-curve should give you an indication of whether the technology is reaching the limits of its efficiency and is therefore ripe for replacement.

4. Historical Analogy

Using historical analogy is a very simple and commonly used method for predicting technology change by comparing the path of development followed by an analogous technology—see Chapter 25, "Historiographical Analysis," for an in-depth treatment of this analytical tool.

5. Scenarios

Scenarios are not strictly predictive; however, they are generally considered a good method for technology forecasting—see Chapter 18 in our previous book, Strategic and Competitive Analysis, for a detailed approach to this technique.

6. Morphological Analysis

Morphological analysis is sometimes referred to as "organized invention." It starts with a goal you wish to achieve. For example, you may be looking to find the optimal method for packaging an object. It involves the systematic gathering of information about all possible technologies that may achieve a particular purpose. For example, you may be looking at packaging technology, and so you would have to consider any possible material that could be used for packaging—cardboard, paper, plastic, fabric, wood, etc. You also need to list the attributes you seek in packaging. For example, you might be considering attributes like durability, flexibility, being lightweight, protective, recyclable, waterproof, and so on.

The information is then displayed in some sort of graphical form—for example, in a list or matrix—that highlights any gaps. The gaps may represent opportunities for developments. The display of the information gathered may also indicate areas with no potential for development at all.

To make this process work for you, you must be prepared to consider all possibilities and not limit yourself to current possibilities. It requires time and patience and may involve some research to ensure you are aware of all the possible technologies relevant to the problem you wish to solve.

7. Relevance Trees

Relevance trees are detailed hierarchies of methods for achieving a particular outcome. This outcome is the question you want answered by your forecast. It might be something like: How can we reduce energy costs for consumers of our appliance?

A relevance tree divides a broad subject/problem into increasingly smaller and more detailed subtopics. Often, relevance trees are arranged to look very much like an organizational chart or family tree, though they may also be represented with more detailed items radiating out from the central subject.

The items at each level of the tree should provide a complete description of the item to which they are joined. Ideally, there should be no overlap between items in the tree; however, this is often difficult to achieve in practice.

The idea behind a relevance tree is to break down a question or problem into issues small enough to be addressed easily. See Figure 22.5 for a diagrammatical representation.

Figure 22.5 Relevance tree model (for pollution control)
Source: Adapted from http://www.wiley.com/college/dec/meredith298298/resources/addtopics/addtopic_s_02m.html.

8. Monitoring

Monitoring is a method for forecasting technology change that does not require a specific question to answer in order to give useful results. There are many forms of monitoring that can predict technology developments. An important one is patent analysis, which is discussed in detail in Chapter 22 of our previous book, Strategic and Competitive Analysis.

Monitoring is often based on careful observation of published research results. Published results will be those emanating from research in public institutions and not competitor firms, so these will tend to be results from very early on in the process of a development (before the research is taken up for commercial pursuit). In some industries, this sort of research may not necessarily be available publicly.

Other sources of information for monitoring include industry publications and trade shows. In some industries, there are associations where individuals with an interest in an invention may publicize their work. Observing developing social phenomena may also give insight into areas where technology development is likely to occur; for example, the wide uptake of cellular phones and the concurrent explosion in use of the Internet led commentators to predict phones with Internet capabilities long before prototypes were built.

The information gathered by your monitoring activities will only be useful if you can analyze your discoveries and find the links between the various observations you make. It is very easy with monitoring to amass huge quantities of information requiring complex and time-consuming analysis. There are software programs available that facilitate this process. Some search through all the data you have stored, looking for links. Others may actively manipulate the data you have found using processes such as network analysis. Network analysis takes your observations and works through multiple combinations to predict a range of scientific capabilities that may be developed.

Using the Information

Whichever technique you use to construct your technology forecast will be irrelevant if you do not use the output of the analysis in some way to enhance your firm's competitive ability. Care should be exercised when acting on the technology forecast. No forecast is ever going to be 100% true, no matter what you pay for it. It is ultimately a statement of probability.

A forecast is limited by the parameters within which it has been made. The predictive value of a technology forecast is lost when those parameters are ignored. An extreme example is a situation where you perform a trend extrapolation over the next five years, but then use the information to support decisions about the next 10 years (you cannot extrapolate the extrapolation and preserve any accuracy at all).

In the end, as an analyst, you should be as aware of the shortcomings of whatever method of forecasting you use, as you are of its strengths. This does not diminish the value of the forecast; rather, it allows you to get the best value you can from the information you have gathered and by ensuring that any decisions made are fully informed by the output of the analysis.

Case Study: Bell Canada and the Delphi Process

Bell Canada is a telecommunications firm with activities in telecommunications operations, research and development, and manufacturing. In the late 1960s, the Business Planning Group (the Group) within Bell noted a range of factors likely to lead to significant medium and long-term changes to the business. These included the merging of computer and communications technologies, new competition due to regulatory changes, emerging visual telecommunications markets, anticipated social change, and increasing costs. The Group developed a Delphi study, which they implemented in 1970 and which predicted a span of 30 years from 1970 to 2000.

The Group divided the business into segments: educational, medical, information systems for business, and residential markets. The study aimed to investigate future applications for Bell Canada's business in these areas.

Before preparing the questionnaire it would submit to its panel of experts, the Group undertook an extensive literature review to explore developments foreshadowed in these areas. The aim was to provide some guidance for the experts. The questionnaire was pretested on a group of experts already available. This allowed the Group to identify and reword some badly expressed questions and redesign the questionnaire to reduce confusion for panel members. Although this step delayed the process, it ensured that results would be clearer.

The education, medicine, and business questionnaires all started by requesting that the panelists give their personal prediction of change to 10 basic values over the next 30 years in North America (for example, would there be a "significant increase," "slight increase," "no change," "slight decrease," or "significant decrease" in "traditionalism," "authoritarianism," "materialism," etc.). This question was to set the mood and put the experts into the right frame of mind for completing the rest of the questionnaire. Other non-technical areas were also put to the panelists. For example, the education study also addressed evolution in school design and the changing role of teachers over the same 30-year period.

The studies then went on to technologies relevant to each area. For example, the education study examined likely time for introduction of computerized library systems, computer-aided instruction systems, and visual display systems across primary, secondary, and post-secondary education levels. Further questions then broke down these hypothetical systems further and asked for a prediction of when they were likely to be in use.

Similar questions about likely timeframes for adoption of hypothetical technology developments were asked in the medical and business studies. The results of these three studies provided clear and useful information.

The residential use study was different in that it focused on future services and not technologies. An early issue with this study was the problem of identifying which "experts" should make the predictions about future adoption of technology for the residential market; for example, was a housewife more qualified by her experience than an industry insider?

The Group solved this problem by conducting two separate Delphi processes with the same questionnaire: one using a panel of housewives and the other using a panel of industry experts. The issues examined by each panel included their opinion of future acceptance of concepts like electronic shopping from home, remote banking, and electronic security for the home.

The Delphi process ran for three years.

The results of the process were then entered into a database where the forecasts from the experts had been indexed for keywords, abstracted, and stored online. Results of other studies—for example, trend analyses—and from internal research were put into the same system.

The information from the Delphi process has been used in combination with other information for a variety of purposes, including to prepare specific service and business proposals and to prepare "environmental outlook reports," which identify future trends that may affect Bell.⁶

FAROUT Summary

Figure 22.6 Technology forecasting FAROUT summary

Future orientation—High. Forecasting is by nature future-focused. The extent of the future focus of a technology forecast will depend on the method used and the questions asked to prompt the forecast.

Accuracy—Low to medium. Technology forecasting does not provide hard data. It can be accurate within the parameters of the forecast, but the information is not accurate in the sense that it does not provide precision. Generally the most accurate information will be about the near future.

Resource efficiency—Medium. Some methods for forecasting are very simple and inexpensive (for example, a straightforward trend analysis); others can be very time-consuming (for example, morphological analysis) and/or quite expensive.

Objectivity—Low to medium. The objectivity of a technology forecast will rest on the nature of the questions asked (do they have assumptions built into them?) and on the people undertaking the process to generate the forecast (are they considering all alternatives necessary for the process to work efficiently?).

Usefulness—Medium to high. Provided care is taken not to treat a technology forecast as a statement of inevitability, and the forecast is interpreted in light of the parameters within which it is made, a technology forecast can be a source of very useful information and strategic opportunity.

Timeliness—Low to medium. Most technology forecasting methods are too timeconsuming to conduct and yield their most accurate information about the near future.

Related Tools and Techniques

• Historiographical analysis
• Patent analysis
• Scenario planning
• SWOT analysis

References

Ascher, W. (1979). Forecasting: An Appraisal for Policymakers and Planners (rev. ed.). Baltimore, MD: Johns Hopkins University Press.

Ayers, R. (1969). Technological Forecasting and Long-Range Planning. New York, NY: McGraw-Hill.

Coates, J.F. (1989). "Forecasting and planning today plus or minus twenty years," Technological Forecasting and Social Change, Vol. 36, Nos. 1–2, August, pp. 15–20.

Coates, V., Faroque, M., Klavans, R., Lapid, K., Linstone, H., Pistorius, C., and A.L. Porter (2001). "The future of technology forecasting," www.tpac.gatech.edu/papers/techforcast-abs.php.

Day, L. (2002). "Delphi research in the corporate environment," in Linstone, H., and M. Turoff (eds.). "The Delphi method: techniques and applications," www.is.njit.edu/pubs/delphibook/ch3c1.html.

Donnelly, Dr. D. (2006). "Forecasting methods: a selective literature review," www.class.uh.edu/MediaFutures/forecasting.html. Accessed April 2006.

Fleisher, C.S., and B.E. Bensoussan (2003). Strategic and Competitive Analysis: Methods and Techniques for Analyzing Business Competition. Upper Saddle River, NJ: Prentice Hall.

Foster, R.N. (1982). "Boosting the payoff from R&D," Research Management, 15(1), pp. 22–27.

Frick, Dr. R.K. (1974). "Operations research and technological forecasting," Air University Review, May–June.

Lawrence, S.R. (2002). "Technology scanning & forecasting," University of Colorado, http://leeds-faculty.colorado.edu/lawrence/mbat6450/docs/schedule.htm.

Lenz, R.C. (1962). "Technological forecasting," ASD-TDR-62-414, Aeronautical Systems Division, Air Force Systems Command.

Mann, D. (1999). "Using S-Curves and trends of evolution in R&D strategy planning," www.triz-journal.com/archives/1999/07/g/index.htm.

Martino, J. (1983). Technological Forecasting for Decision Making. New York, NY: Elsevier Science Publishing Company.

Meredith, J., and S. Mantel (2000). Project Management: A Managerial Approach, 4th edition. New York, NY: John Wiley & Sons.

Millett, S.M., and E.J. Honton (1991). A Manager's Guide to Technology Forecasting and Strategy Analysis Methods. Columbus, OH: Batelle Press.

Oxford Dictionary. www.askoxford.com/concise_oed.

Schnaars, S. (1989). Megamistakes: Forecasting and the Myth of Rapid Ttechnological Change. New York, NY: The Free Press.

Web sites:

http://hops.wharton.upenn.edu/forecast

http://www.wiley.com/college/dec/meredith298298/resources/addtopics/addtopic_s_02m.html

Endnotes

¹ Martino (1983).

² Meredith and Mantel (2000).

³ See Fleisher and Bensoussan (2003), Chapter 22.

⁴ See Fleisher and Bensoussan (2003), Chapter 18.

⁵ See Fleisher and Bensoussan (2003), Chapter 24.

⁶ Adapted from Day, L. (2002).