Technological Forecasting

1
Technological Forecasting

• Henry C. Co
• Technology and Operations Management,
• California Polytechnic and State University

2
What is Technological Forecasting?

• Technology forecasting is forecasting the future
  characteristics of useful technological machines,
  procedures or techniques.
• Items which depend on popular tastes rather than
  on technological capability are excluded. Thus
  commodities, services or techniques intended for
  luxury or amusement are excluded from the domain
  of technological forecasting.
• The forecast does not have to state how these
  characteristics will be achieved.

3
What Characteristics?

• Levels of technical performance, like speed of a
  military aircraft, the power in watts of a
  particular future engine, the accuracy or
  precision of a measuring instrument, the number
  of transistors in a chip in the year 2015, etc.
• Dates and probabilities of breakthrough events,
  technical parameter trends, technology
  substitution rates, technological impacts, and to
  some extent, market growth trends.

4
Elements of Technological Forecasting

• Technology being forecasted.
• Time of the forecast a single point, or a time
  span.
• Statement of functional capability a
  quantitative measure of its ability to carry out
  the function.
• Statement of
• Probability of achieving a given level of
  functional capability by a certain time or
• Probability distribution over the levels that
  might be achieved by a specific time.

5
What are we forecasting?

• A specific technical approach, or a more general
  technology?

6
Whats the difference?

• Specific technical approachmeans of solving a
  problem/performing a particular function.
• For example, Piston engines and jet engines are
  two different technical approaches of the
  technology of powering aircraft
• Incandescent lamps, fluorescent lamps, and arc
  lights are different technical approaches to the
  technology of providing illumination.
• A technical approach may be further subdivided.
  e.g., jet engines can be divided into turbojets
  and turbofans.

7
Functional vs. Technical Parameter

• Functional parameters directly measure the
  extent to which the technology satisfies the
  users needs (speed, power, etc.)
• Designers adjust combination of technical
  parameters (e.g., turbine inlet temperature and
  compression ratio) to achieve the functional
  parameters desired by the engine user.

8
Why Forecast Technology?
Performance
Physical limit of technology
Effort (funds)
Foster, Innovation The Attackers Advantage,
Summit Books, 1986
9
Why Forecast Technology?

• To maximize gain from events external to the
  organization.
• To maximize gain from events that are the result
  of actions taken by the organization.
• To minimize loss associated with uncontrollable
  events external to the organization.
• To offset the actions of competitive or hostile
  organizations.

10
Why Forecast Technology?

• For purposes of production and/or inventory
  control.
• For facilities and for capital planning.
• To assure adequate staffing.
• To develop administrative plans/policy internal
  to an organization (e.g., personnel or budget).
• To develop policies that apply to people who are
  not part of the organization.

11
For Decision Making

• Identifies limits beyond which it is not possible
  to go.
• Establishes feasible rates of progress, so that
  the plan can be made to take full advantage of
  such rates the plan does not demand an
  impossible rate of progress.
• Describes the alternatives that can be chosen.
• Indicates possibilities that might be achieved if
  desired.
• Provides a reference standard for the plan. The
  plan can thus be compared with the forecast at
  any later time to determine whether it can still
  be fulfilled or whether, because of changes in
  the forecast, the plan must be revised.
• Furnishes warning signals, which can alert the
  decision maker that it will not be possible to
  continue the present activities.

12
Technological Forecast Is Self-altering

• Weather forecast must be correct if it is to be
  useful.
• Technological forecast is self-altering.
• A self-altering forecast is one that, by virtue
  of having been made, alters the outcome of the
  situation.
• Suppose someone forecasts an undesirable
  situation. Then suppose a decision-maker accepts
  the forecast and acts to prevent the undesirable
  situation. Clearly the forecast did not come
  true. Was it a bad forecast?
• It is even more important that forecasters
  educate forecast users to the idea that the
  goodness of a forecast lies in its utility for
  making better decisions and not in whether it
  eventually comes true.

13
Contour Map of the Future

• Future fans out as a wedge-shaped terrain of
  peaks/valleys of threats/opportunities
• Probability of following any one given pathways
  into the future is small, but the sum of the
  probabilities of all the different discrete
  pathways through the terrain 1.
• Forecasters job is to map out the contours
  (threats and opportunities) of the futures
  terrain and show the potential routes through it
  so the decision maker can judge the best path.
• Forecasters Dilemma the finer the details used
  to describe the pathway through the terrain, the
  lower the probability of that exact pathway being
  followed and that particular terrain being
  traversed as the future unfolds.

(after Porter et al, 1991)
14
Process and Philosophy
There is no such thing as a value-free forecast
Its influence starts with where and how we search
for and select our input data and continues on
through how we analyze and interpret the results.
15
Methods of Forecasting

• Growth curves and Extrapolation
• Leading indicators
• Causal models
• Probabilistic models
• Environmental Monitoring

16
Extrapolation

• Assumption Time series data from the past
  contains all the information needed to forecast
  the future.
• The forecaster extends a pattern found by
  analyzing past time series data.
• For example A technological forecaster who was
  attempting to forecast aircraft speed would
  obtain a time series of aircraft speed records,
  find a pattern (trend), and extend to the future
  to obtain a forecast.

17
Trajectory of Tech Innovation
Physical limit of technology
Performance
Effort (funds)
Technological performance often follows an
S-shaped curve
Foster, Innovation The Attackers Advantage,
Summit Books, 1986
18
The Pearl-Curve (Excel)
19
The Pearl-Curve (Excel)
20
Leading Indicators

• Assumption The time series of interest shows the
  same behavior as another time series (the leading
  indicator), but with a known time lag. Thus, what
  the leading indicator is doing today will be
  matched by the time series of interest at a
  specific time in the future.
• The forecaster uses one time series to obtain
  information about the future behavior of another
  time series.
• For example A weather forecaster uses turning
  point in the time series of barometric pressure
  to forecast a future turning point in the amount
  of precipitation.

21
Causal Models

• Assumption The cause-effect linkages in the
  topic of interest are known and can be expressed
  mathematically or in some similar fashion (e.g.,
  a mathematical model).
• Incorporates information about cause and effect
  relationship, involving some fundamental laws in
  physics.
• For example A forecast of solar eclipse is based
  on a causal relationship, involving some
  fundamental laws of physics.

22
Probabilistic Models

• Instead of producing a single-valued forecast,
  probabilistic models produce a probability
  distribution over a range of possible values.
• Example The probability of rain tomorrow may be
  stated as, for instance, 30. This means that
  over the range of possible outcomes, rain and
  no-rain, the associated probabilities are 30 and
  70, respectively.

23
Environmental Monitoring

• Forecasts based on trends or growth curves
  require continuity between the past and the
  future.
• Forecasts based on causal models require the
  consistent operation of the causal factors.
• Environmental Monitoring of precursor events made
  it possible to forecast the eventual development
  of breakthrough technologies.

24
Forecasting a technological breakthrough requires
that precursor events be identified and used to
provide advance warning.
(after Fahey Narayanan, 1987)
25
Process of Monitoring

• Collection
• Screening
• Evaluating
• Threshold-setting

26
Collection

• Aircraft engine company (1930) Concerned about
  possible threats to its business of reciprocating
  aircraft engines.
• The firm's technological forecasters track
  patents granted in the field of aircraft
  propulsion.
• Patent granted in 1930 to Flying Officer Frank
  Whittle (Royal Air Force) for an aircraft engine
  based on the jet principle.
• Air is drawn in through a turbine compressor,
  fuel is burned in the compressed air, and the
  combustion gases are used to drive a turbine,
  which in turn drives the compressor, while
  providing some net thrust to propel the aircraft.
• This important signal must be screened for
  significance.

27
Screening

• The jet engine patent is clearly significant to
  an aircraft engine company. It is potentially a
  disruptive technology. The forecaster should
  search further for past signals.
• 1910 Henri Coanda proposed a jet propulsion
  system in which the compressor would be driven by
  a reciprocating engine instead of by an exhaust
  turbine (ramjet).
• 1913 French Engineer Rene Lorin proposed a jet
  engine in which the compression is derived
  entirely from the aircrafts forward velocity,
  eliminating the need for a compressor
  (turboprop).
• 1921 Maxine Guillaume received a French patent
  on a jet engine with a turbine-driven compressor
  similar to Whittles design.
• 1929 A. A. Griffith of the Royal Aircraft
  Establishment proposed that a turbine engine be
  used to drive a propeller for providing aircraft
  propulsion (turbojet).

28
Evaluating

• What does this mean to my organization?
• If it represents the start of a trend or pattern,
  would it affect our mission?
• Would it make a product obsolete?
• Would it alter a production process?
• Would it have an impact on a customer? A
  supplier?

29
Whittles Patent

• For Whittles engine to work, a compression ratio
  of 4 to 1 and a compressor efficiency of 75
  would be required. The turbine blades would have
  to withstand a temperature of 1500?F. Thus a
  forecaster would be interested in seeking
  information on these parameters.
• A 1923 report published by Dr. Edgar A.
  Buckingham (National Bureau of Standards) showed
  that only at speeds above 500 mph would the jet
  engine be competitive in fuel economy. Hence the
  forecaster would also be interested in tracking
  aircraft speed.

30
Threshold-Setting

• As evaluation continues, the evidence for one or
  more hypothesis will become stronger and
  stronger. When the confirming signals show that
  the hypothesis has exceeded its threshold, it is
  time to make a breakthrough forecast.
• Thresholds passed by 1938 Whittles engine
  requires compression ratio of 4 to 1 and
  compressor efficiency of 75. Turbine blades
  would have to withstand a temperature of 1500?F.
• In 1931, compression ratio and compressor
  efficiency were 21 and 65, respectively.
• By 1935, these had reached 2.51 and 65.

31

• In 1935, Hans von Ohain obtained a German patent
  on a turbojet engine similar to Whittles
  received support from the Heinkel company for
  development of a jet engine.
• In 1936, Whittle founded Power Jets Ltd. To
  develop an engine according to his design.
• In 1938, the U.S. Army Air Corps laboratories at
  Wright Field (Dayton, OH) began a 5 year program
  of development of gas turbines for jet engines.
  NACA began a program of compressor development.
  RAE began work on a turbocompressor based on
  Griffiths 1929 design.

32

• Hans von Ohains jet engine achieved flight in
  1939.
• Whittles engine flew in 1941.
• In 1940, the Caproni-Campini CC-2 flew with a
  Coanda-type jet engine.
• In 1942, a U.S. aircraft flew using a jet engine
  developed by General Electric. That same year saw
  the flight of German jet aircraft that was no
  longer experimental but a combat aircraft.

33
Technological Breakthrough
34
Forecasting a Breakthrough

• Breakthroughs in technology do not come as bolts
  from the blue.
• Breakthroughs are the end result of a chain, or
  even a network of precursor events, and these
  events give warning of a breakthrough is coming.
• Forecasting a technological breakthrough requires
  that precursor events be identified and used to
  provide advanced warnings. The monitoring process
  is designed to help the forecaster answer two
  question
• Which events are precursors?
• What do the precursors do?
• See Martino, J. P., Using Precursor as Leading
  Indicators of Technological Change, 32 341-360
  (1987).

35
Forecasting a Breakthrough

• Involves a systematic search for these precursor,
  coupled with an evaluation of the significance of
  the precursor.
• The forecaster seeking advanced warning of a
  breakthrough must search all the relevant sectors
  of the environment in order not to miss important
  signals of coming breakthroughs.
• The signals found must be synthesized into
  possible patterns of change, and the forecaster
  should continue to search for additional signals
  suggested by the hypothesized patterns.
• It is important to search for both confirming and
  disconfirming signals.

36
Precursor Events

• There are many precursor events leading to a
  breakthrough.
• For instance, there are many precursor events
  between the unpredictable scientific breakthrough
  of 1905 and the eventual commercial use of atomic
  power in 1956.
• Not all these precursor events provided positive
  signals.
• Some, such as the impracticality of atomic energy
  plants using particle accelerators, were false
  negative signals.
• Some, such as the possibility of fusing light
  atoms into moderately heavy ones, pointed in the
  wrong direction.
• Nevertheless, atomic energy was not an unheralded
  event.

37
Precursors to Commercial Use of Atomic Power