Chapter 7

Technological Diffusion and Disruption

Advanced Technology Roadmap Architecture (ATRA)

Inputs
Steps Outputs
Strategic Drivers for Technology
+10y Technology
FOMjj

1. Where are we today? Roadmaps

L1 Products and Missions +5y

L2 Technologies Technology State of the Art and Organization Figures of Merit (FOM)
Competitive Benchmarking Current State of the Art (SOA)
Competitor 1 Technology Trends dFOM/dt
Technology Systems Modeling Competitor 2

Tech Pul
Pull Today FOMi
2. Where could we go?
Dependency Structure Matrix

L1
+10y
Scenario-based
Technology Systems Modeling and Scenario A Technology Valuation
FOMj

Trends over Time

Technology +5y Design Reference Missions
Scenario B
Projects Future Scenarios
? T h l
Technology V
Valuation
l ti
3. Where should we go? Vector Charts
L2
Scenario Analysis and FOMi
Technology Valuation
Tech Push Technology Investment
Efficient Frontier
E[NPV] - Return

Technology Scouting 4. Where we are going! Pareto-optimal set of technology

Knowledge Management Technology investment portfolios
Technology Portfolio Valuation, Portfolio Recommended Technology Portfolio
Intellectual Property Analytics Optimization and Selection Technology (Expected NPV and Risk)
Projects
σ[NPV] - Risk
Foundations C
Cases
7
Definitions History Nature Ecosystems The Future Case 1 Case 2 Case 3 Case 4
What is Milestones of Technology, Nature Technology Diffusion, Is there a Deep Space DNA
Singularity ? Automobiles Aircraft
Technology? Technology and Humans Infusion and Industry Network Sequencing

© Springer Nature Switzerland AG 2022 183

O. L. de Weck, Technology Roadmapping and Development,
[Link]
184 7 Technological Diffusion and Disruption

7.1 Technology Adoption and Diffusion

Once invented and “launched” a technology will initially have few followers or adopt-
ers. This is normal as the technology is often initially unknown, except to the inventors
themselves, and some opinion leaders who may become “early adopters.” Some of the
earliest work on the topic of “Diffusion of Innovations” was done by Everett M. Rogers1
in his landmark book “Diffusion of Innovations” first published in 1962 (Rogers,
2003).2 The book was based on his 1957 doctoral dissertation, which was on the topic
of adoption of agricultural innovations in the rural community of Collins, Iowa.
As a social scientist, he interviewed many of the 148 farmers in that community
to better understand what prompted them to adopt early or delay adopting agricul-
tural innovations. In fact, the term “early adopters” was coined by him.
Rogers grew up on a rural farm in Iowa and witnessed his father, who was a
farmer, readily adopt electro-mechanical innovations (such as the tractor) but be
much slower when it came to bio-chemical innovations such as hybrid corn seeds,
or 2,4-D weed spray. This sparked his interest in how individuals decide if and when
to adopt innovations, such as new technologies. His study of diffusion of innova-
tions was not confined to the adoption of new technologies per se. He also studied
other “innovations” or policy interventions such as practices to slow the spread of
HIV/AIDS, family planning, and nutrition. Rogers defines innovation as follows:

✦ Definition
Innovation is defined as an idea, practice, or object that is perceived as new by
an individual or other unit of adoption. An innovation presents an individual
or an organization with a new alternative or alternatives, as well as new means
of solving problems.

The main contribution of Rogers work is the elaboration of a general model of

diffusion of innovations which is independent of the nature of the innovation itself
(technological or other) but does assume differences in individual’s willingness to
adopt new solutions. It is based on the Gaussian distribution (the well-known “Bell
curve”) shown in blue in Fig. 7.1. Integrating under the blue curve which shows the
number of adopters that adopt early or late yields the yellow cumulative “S-curve”
which is based on the logistics function, which was already discussed in Chap. 4.
The different categories of technology adopters and assumed fractions of the
population of potential adopters are: innovators 2.5%, early adopters 13.5%, early
majority 34%, late majority 34%, and laggards 16%. These are defined by the -3σ,
-2σ, -1σ, +1σ, and +2σ (extended to infinity, i.e., those that never adopt) boundaries
of the underlying Gaussian distribution. Rogers found in his interviews that
innovators – the first individuals in a system to adopt an innovation – tended to be
individuals who travel a lot, read widely, and have a “cosmopolitan” mindset.

1
The first edition appeared in 1962, while the fifth and latest edition was published in 2003.
2
Some point out that even though Rogers is better known, that it is really Zvi Griliches, a Harvard econo-
mist who should deserve the credit for being the first to rigorously study technology adoption (1957).
7.1 Technology Adoption and Diffusion 185

Fig. 7.1 The diffusion of innovations according to Rogers. With successive groups of consumers
adopting a new technology (shown in blue), its market share (yellow) will eventually saturate.
(Source: [Link] This
assumes total substitution and fixed market size

This general diffusion model has had an enormous impact3 and is generally still
viewed as a valid way to think about technological adoption as a universal process
of social change.
Is there empirical evidence that this diffusion of innovations model is correct
when it comes to the adoption of technologies? Fig. 7.2 highlights some of the origi-
nal research done by Rogers on the adoption of hybrid seed corn in Iowa in the
1930s and 1940s. The facts of whether and when an individual farmer became an
adopter of hybrid seed corn were painstakingly established mainly through personal
interviews in the community. The cumulative curve clearly features an S-shape,
albeit an asymmetric one.
Figure 7.2 shows, it took a full decade from when the first farmer adopted hybrid
seed corn in 1927, until the peak of adoption in a single year which was 1937. As
mentioned, Rogers’ own father was not an early adopter. During the Iowa drought
of 1936, however, while the hybrid seed corn stood tall on the neighbor’s farm, the
crop on the Rogers’ farm wilted. Rogers’ father was finally convinced. The 1936
drought and peak in 1937 help explain the location of the year with the largest num-
ber of new adopters (1937). This shows that the process of deciding whether or not
and when to adopt the new technology is an individual choice. The results of inno-
vation take time to manifest themselves since the skeptics want to first see “proof”
of the value of the new technology. In agriculture, for example, one has to wait for
one or more annual growing seasons to see the net results of an innovation. Also,
interpersonal contacts and opinions shared across one’s personal or professional

3
In the early 2000s, Diffusion of Innovations was the second most cited text in the social sciences.

Licensed to ([Link]@[Link])
186 7 Technological Diffusion and Disruption

Fig. 7.2 The number of new adopters each year, and the cumulative number of adopters, of hybrid
seed corn in two Iowa farming communities (Source: Rogers 1962)

network of peers appear to play a large role. Consider, for example, Fig. 7.3, which
shows a social network of early adopters reconstructed by Rogers through his
interviews.
Adopter No.1 heard about the innovation from an agricultural scientist (middle
right) and first tried the new weed spray in 1948. Then in 1950 (2 years later!),
farmer No.2, who knew No.1, also adopted the weed spray. This farmer then became
an opinion leader for eight other farmers who also became adopters between 1951
and 1956. Clearly, the social network and credibility of farmer No.2 played a large
role in the diffusion of this technology in this particular community.

➽ Discussion
Can you think of an example where you personally were trying to decide
whether or not to adopt a new technology or wait until later? Who or what
influenced your decision?
How would you classify yourself in terms of the groups shown in Fig. 7.1?
7.1 Technology Adoption and Diffusion 187

Fig. 7.3 The diffusion of new weed spray in an Iowa farm neighborhood. Note the direction of the
dashed arrows is from the later adopter to the earlier adopter. For example, farmer No.10 (1954)
followed No.3 (1951), who followed No.2 (1950), who in turn followed No.1 (1948), who was the
original adopter
188 7 Technological Diffusion and Disruption

Fig. 7.4 Layout of the QWERTY (top) and Dvorak (bottom) keyboards

An important part of understanding the diffusion model is the role of (epistemic)

uncertainty.4 Individuals or organizations, when confronted with the decision to
hold on to an existing solution versus adopting a new, lesser known one, are facing
a decision under uncertainty. In order to reduce this uncertainty (the downside of
this uncertainty is what we call “risk” and the upside is “opportunity”), potential
adopters seek information from any possible source such as subjective opinions of
peers, publications in the trade press, and increasingly the Internet. As Rogers
stated: “This information exchange about a new idea occurs through a convergence
process involving interpersonal networks. The diffusion of innovations is essen-
tially a social process in which subjectively perceived information about a new idea
is communicated from person to person.” In this sense, manifested technological
diffusion is essentially the cumulative effect of thousands or millions of individual
decisions. The main decision alternatives are: adopt now – wait – don’t adopt.
An important point is that even when a specific technology is shown to be benefi-
cial and superior when it comes to certain figures of merit (FOMs), there is no
guarantee that a particular innovation or technology will be broadly adopted. A clas-
sic example that has been claimed by Paul David (1985) is the layout of the com-
puter keyboard, see Fig. 7.4.5
While the Dvorak keyboard is optimized for typing efficiency in the English
language and takes into account the frequency of words in the alphabet (e.g., notice

4
Epistemic uncertainty is that uncertainty where the information is unknown to the decision maker, but
the facts are already established and knowable. This is in contrast to aleatoric uncertainty where the
facts are not yet established and are subject to a random stochastic process that unfolds in the future.
5
However, in a more recent article by Liebowitz and Margolis (1990), the claim of the superiority of
the Dvorak keyboard over QWERTY has been severely challenged, and some would say debunked.
7.1 Technology Adoption and Diffusion 189

the prominent locations of the letters “E” and “T” in the home row), it was never
widely adopted. The QWERTY keyboard, on the other hand, which was designed
more than a century ago in 1873 to slow down typists so as to prevent the jamming
of neighboring keys on a mechanical typewriter was never displaced. More on the
factors that can promote or hinder technological diffusion and disruption will be
discussed in a later section. In summary: successful technological diffusion is not
inevitable. And, invention and diffusion are distinct processes that must be consid-
ered in their own right.

⇨ Exercise 7.1
Think of a technology that you have heard about that may have been superior
based on technical merits, but that was ultimately not adopted at a wide scale.
What may have contributed to it not being adopted?

According to Rogers, the four key ingredients needed in the successful diffusion
of an innovation are: (1) the innovation itself – for example, a new technology; (2)
communications about the new innovation through one or more channels; (3) time;
and (4) a social system through which information about the innovation travels and
which will (or not) adopt the new technology over time. The absence or lack of any
of these four ingredients can doom the diffusion process.
In his fifth (and last) edition of Diffusion of Innovations, which was published in
2003,6 Rogers also looked at the diffusion of new communications technologies
such as mobile phones and the Internet. Fig. 7.5 shows the estimated adoption curve
for cellular (mobile) telephones in Finland between 1981 and 2002.
It is interesting to note that while Finland was a pioneer in mobile radio com-
munications (e.g., for some time Nokia was a dominant player in that industry) even
after 20 years there were still over 20% of the population who had not adopted
mobile phones. The reasons for late or no adoption can be varied, including lack of
financial resources, mistrust, lack of perceived need, or simply unawareness, which
can be linked to a lack of communication and isolation of individuals. There is gen-
erally an assumption that older adults (those over age 65) adopt technologies at a
lower rate than younger adults or adolescents. We will examine this aspect in
Chap. 21.
Making individual technology adoption decisions is based on several factors,
including subjective information from peers about the effectiveness (and, therefore,
value) of new technologies. The Internet has had a large effect on technology adop-
tion due to its peer-rating systems such as the now well-established five-star (*****)
ratings on many sites (such as [Link]). In a sense, the Internet has not only
itself been adopted at a very fast rate (see Fig. 7.6), but access to the Internet has
indirectly acted as an accelerator of technological adoption.
After the first computer network, ARPANET, was established by the US
Department of Defense in 1969, it took almost 20 years for the Internet, as we know

6
Everett M. Rogers passed away on October 21, 2004, in Albuquerque, New Mexico.
190 7 Technological Diffusion and Disruption

Fig. 7.5 Rate of adoption of mobile telephones in Finland (1981–2002)

it today, to emerge. Keys to broad diffusion of the Internet were complementary

innovations such as the html protocol (the World Wide Web), web browsers, and the
appearance of commercial applications online.
An important takeaway from the study of diffusion of innovations is that the
innovation, whether it is a new technology or not, will not be adopted “on its own”
or automatically as the example of QWERTY shows. A complex social system is
required along with careful and targeted communication in order to lead to success-
ful adoption. An interesting distinction is that between a centralized and a decentral-
ized innovation diffusion system, see Fig. 7.7.
In the centralized approach, the innovation is deliberately incubated in the
Research and Development (R&D) part of an organization. While (hopefully) based
on a real or latent demand, the innovation is then “pushed” by a change agent toward
the innovators and early adopters, some of whom may act as opinion leaders. An
example of a commercial version of a centralized diffusion system is the highly suc-
cessful launch and adoption of the iPhone by Apple. This product has not only
found many millions of buyers worldwide, it has also cannibalized – we may say
disrupted – the sales of classic mobile phones (such as the still existing but almost
extinct flip phones) and created a whole new software industry for Apps.
Another example of centralized diffusion of technological innovation, where the
Change Agent was a national government, is the adoption of nuclear power for elec-
tricity generation in France (Doufene et al. 2019). Fig. 7.8 shows the generation
capacity of nuclear power in France between 1945 and 2012, relative to its
7.1 Technology Adoption and Diffusion 191

Fig. 7.6 Cumulative rate of adoption of the Internet worldwide (Rogers 2003)

competitors: renewables and fossil fuels. The technological change was triggered
by the oil crisis of 1973 and a strategic decision by the French government (due in
large part to the lack of domestic sources of oil and gas) to build a large number of
nuclear power plants and to develop the associated supporting industry. This high
fraction of nuclear power is one of the reasons for the potential of successful adop-
tion of electric vehicles (see Chap. 6) in France.
192 7 Technological Diffusion and Disruption

Fig. 7.7 Centralized (top) and decentralized (bottom) diffusion systems (Rogers 2003)

Fig. 7.8 Market share of electricity generation technologies in France over time. (Source: IDCH
(2001), Varon (1947), INSEE database (2014))
7.1 Technology Adoption and Diffusion 193

It is surprisingly straightforward to create a mathematical model of diffusion of

innovations and implement it in a software simulation. Here, we develop such a
model using the so-called agent-based modeling (ABM) approach.
The simplest kind of these models assumes a normally distributed (Gaussian)
population of potential adopters – each individual being an “agent” – who at the
start, t=0, are all using the existing technology, that is, their new technology adop-
tion “flag” is set to a=0. At t=0, we assume that one agent out of a population of N
individuals will be the first adopter. They will then interact with C other randomly
chosen individuals in the existing population of nonadopters at each time period. It
is implicitly assumed that each interaction will either lead the other individuals
(who are not yet adopters) to either adopt (or not) the technology. This is mathemat-
ically done by comparing the individuals’ initial predisposition to adopt, that is, the
normally distributed variable n against a uniform random number nr on the interval
[0,1], see distribution in Fig. 7.1. If nr <= n, then the individual will adopt the tech-
nology, otherwise they will not. If they do adopt the new technology, their adoption
flag is set from a=0 to a=1 and they join the pool of adopters who now get to inter-
act with nonadopters in the next time step.
We set some simple simulation parameters, such as:
N = 10,000; population size
C = 1; number of contacts of each adopter with a randomly chosen nonadopter per
time period
T = 100; number of time periods
A result from an instance of such a simulation is shown in Fig. 7.9.
We see that initially the adoption rate is relatively low, but “take-off” – as Rogers
called it – occurs after about 10 time periods and we reach the midpoint (50% of the
population have adopted) at 21 time periods. After t=40, nearly 100% of the popula-
tion have adopted. Technically, since the Gaussian distribution is unbounded, we
can never reach exactly 100% and some individuals will never adopt.
These kinds of diffusion models are also used to study the diffusion of other
phenomena, besides technology, such as the spreading of diseases or rumors.
The (Matlab) code for this simulation is given in the Appendix.
It is important to note that this is a stochastic model, meaning the result is some-
what different, every time the simulation is run, since both the initial spread of the
potential adopters and their bilateral contacts are randomized. For example, execut-
ing this simulation 10 times yields results for the midpoint between 21 and 28 time
periods. The midpoint of diffusion, when 50% of the population have adopted the
technology, is itself a normally distributed random variable.
With such mathematical models of diffusion, one can now experiment with dif-
ferent effects that could impact the rate of diffusion. For example, allowing five
contacts of each adopter per time period (instead of only one) shortens the time to
reach the midpoint by about a factor of 3 (to t=8 time periods).
This mimics the impact of more frequent communication and marketing.
Increasing the population size to N=50,000 lengthens the time to reach the midpoint
by about a third to t=30. One can imagine that there are several important
194 7 Technological Diffusion and Disruption

Fig. 7.9 Agent-based model simulation of technology adoption over time (N=104)

modifications to such a model that would reflect the particular innovation or social
system one is interested in. Calibration and validation of such diffusion models
against real-world data is also an important and tricky issue.
Newer topics in the diffusion of innovations research are diffusion over social
networks, where contacts between adopters and nonadopters are not random – as in
the ABM above – but follow along the edges of an established social network. Also,
as Rogers points out, the decision to adopt is not a one-time event and multiple
exposures to a new technology or idea may be necessary for any individual before
coming up with a definitive decision to adopt or not.

⇨ Exercise 7.2
Trace the diffusion of a specific technological innovation that interests you.
Instead of doing a web search for a preexisting figure, gather your own origi-
nal information, do some background reading, and possibly perform a couple
of interviews with subject matter experts (SMEs) on the topic. Produce a fig-
ure such as Fig. 7.6 from primary data or a simulation such as in Fig. 7.9 and
provide a narrative explaining your results.
7.2 Nonadoption of New Technologies 195

Finally, it is important not to confuse the S-curves discussed here, which pertain
to the adoption and diffusion of technologies over time in a finite size population,
with the (potential) saturation of the performance (or other FOMs) of a technology
due to technical, physical, or other constraints in the system. Some scholars dispute
the existence of FOM-based S-curves (see Chap. 4), while the S-curves and satura-
tion effects in the diffusion of technologies and innovations are well accepted.
Moreover, if the underlying population of potential adopters itself is growing, then
N(t) itself is a function of time and full saturation may not be achieved, or it may be
delayed.

7.2 Nonadoption of New Technologies

The work of Rogers (1962) and others like Griliches (1957) in technology diffusion
brings up some interesting fundamental questions regarding the adoption of tech-
nologies, such as the following:
• Does new technology eventually get adopted by 100% of the population or adop-
tion units, or is this not the case?
• Has the speed of technological diffusion changed over time?
• If some older technologies survive and the adoption of new technologies is not
total, but saturation occurs before reaching 100%, what governs the relative mar-
ket share at equilibrium?
• How do we properly obtain and validate data – besides knocking on doors and
interviewing individual adopters or nonadopters – regarding technology ­adoption
and diffusion and how do we interpret it in a local, regional, national, and global
context?
• How are adoption rates of technologies and their FOM evolution (see Chap. 4)
coupled?
A chart like the one shown in Fig. 7.10 may provide some initial answers.
The different curves in Fig. 7.10 each represent the % of the US population hav-
ing adopted a certain technology over time. According to this chart, only 10% had a
telephone by 1905 (presumably those in a more affluent socioeconomic position),
and it took until about 1945 for 50% of the population to have a (wired!) telephone
in their homes. On the other hand, the Internet and many of the more recent infor-
mation technologies (mostly shown in black, orange, and blue) reached the 50%
mark within a decade or even faster.
So, without a doubt – at least for consumer-type technologies – the adoption has
accelerated considerably in the late twentieth century and early twenty-first century.
There are some important nuances, however, which are often glossed over:
• Diffusion can be nonmonotonic. As can be seen in the early 1930s, the adoption
rate of telephones and cars receded by about 10%. This is presumably due to the
rates of poverty resulting from the Great Depression. The adoption of air travel

Licensed to ([Link]@[Link])
196 7 Technological Diffusion and Disruption

Fig. 7.10 Adoption rates over time for different technologies: 1900-present day in the United
States. (Source: Nicholas Felton, The New York Times)

receded slightly in the early 1970s, probably due to the 1973 oil crisis and ram-
pant inflation.
• Diffusion can saturate at less than 100%. Not every individual or household
owns a color television or credit cards, some people refrain from social media,
and so forth. Technology adoption discussions often give the false impression
that adoption eventually always reaches 100%. This is particularly true when we
take a global perspective. Many households in places like rural India, Africa,
Central America, Southeast Asia, etc. do not have refrigerators, a central water
supply system, and so forth. Even though mobile cellular service exists in most
populated places on Earth today.7
• The data in such diffusion charts may be suspect. It may be questionable and not
collected and validated with scientifically sound methods. Are the data based on
government statistics, user surveys, guesses, etc.? This issue of data validity is a
major concern of and for the technology diffusion research community.
What is often neglected to take into account, when reporting figures on techno-
logical adoption, is the fact that there exist – globally speaking – several communi-
ties that never adopt new technologies, whether by choice or due to a lack of
communication (awareness), lack of economic ability to pay, religious conviction,
cultural incompatibility, or misalignment with the needs of said population.

7
And with the ongoing launch of new Low Earth Orbit (LEO) satellite constellations such as
OneWeb, Starlink, Kuiper, and others, there will soon be 100% global coverage for mobile broad-
band internet access.
7.2 Nonadoption of New Technologies 197

There are essentially four kinds of populations that do not adopt what we might
call “modern” technologies such as the ones shown in Fig. 7.10. The reasons are
varied but include geographical isolation, religious conviction, and a disaffection
with modern society as we know it today.
We briefly describe these four cases and show these situations in Fig. 7.11:
• Indigenous island populations that have been geographically isolated from main-
stream technologically based society, and are also referred to as “uncontacted
peoples.” An example of such a population that has recently been in the news due
to the killing of an unauthorized intruder are the so-called Sentinelese who live
in the Bay of Bengal near India. Photography is prohibited by the Indian authori-
ties and so we only provide a map of their location (see Fig. 7.11a).
• The Old Order Amish, who are mainly centered around the US Midwest and
Lancaster County, Pennsylvania, in particular, eschew the use of modern

Fig. 7.11 (a) Upper left: location of North Sentinel Island in the Bay of Bengal, India; (b) upper
right: a horse-drawn carriage carrying an Old Order Amish family; c) lower left: a mud hut locally
known as a “kaypay” in Haiti (Source: [Link]
suffered-­in-­poverty/) and d) lower right: the abandoned village of O Penso in Northwestern Spain
(Source:[Link]
for-­sale-­and-­theyre-­going-­cheap)
198 7 Technological Diffusion and Disruption

t­ echnology such as electrical machines, automobiles, birth control, and even but-
tons (they use hooks and eyes instead). They are believed to number about
250,000 people today and are growing in numbers given high birth rates (a typi-
cal Old Order Amish family has about 6–7 children on average). The rules of
technology nonadoption are strictly enforced. A classic image of an Amish fam-
ily riding in its horse-drawn carriage is in Fig. 7.11b.
• As mentioned earlier, there are poor populations across the globe, many of them
in the Southern Hemisphere, who simply cannot afford new technologies, even
though their adoption would benefit them, for example, in the area of water and
sanitation. In the Western Hemisphere, the nation of Haiti is the poorest in terms
of GDP/capita and as a result technology adoption rates, particularly in rural
areas, are rather low. As described by Rogers (the case of boiling water in the
Peruvian village of Los Molinas), in some cases, there are also cultural traditions
or superstitions that prevent the adoption of new and technologically enabled
practices. More recently, the United Nations has created the so-called United
Nations Technology Innovation Labs (UNTIL) as a mechanism to promote the
achievement of its UN goals via technological innovation.
• The fourth and final group of technology nonadopters are those who were tech-
nology adopters for a while (e.g., they lived in the larger cities of North America,
Western Europe, or East Asia) and decided for one reason or another to withdraw
from modern society. We may call this group the technology dropouts. These
individuals (in rare cases, entire families) have become disenchanted with mod-
ern technology for different reasons such as their negative impact on the environ-
ment (e.g., the organic movement in agriculture is related to this), their conviction
that technologies will inevitably lead to a “doomsday” for humanity and our
inevitable destruction as a species, or more simply because of unemployment
and poverty. An interesting recent example is a group of dropouts in Northern
Spain8 or individuals living in forests around affluent European cities like Bern,
Switzerland.9
The future of these communities is highly uncertain.
Will they eventually die out and concomitant with urbanization will eventually
100% of the world’s population live in high-rise buildings in mega-cities, while eat-
ing food produced from urban farms or automated rural farms tended by robots? Or
will there be a strong enough “back to nature” movement of people who
deliberately shun technology and go back to prehistoric times (see Chap. 2)? Will
biotechnology find a way to combine the best of modern technology with nature to
reach a new equilibrium (see Chaps. 3 and 22)? At this point there is no way to
know, but we may of course speculate. In conclusion, it should be acknowledged

8
In Northern Spain entire abandoned and vacant villages are up for sale: [Link]
tions/parallels/2015/08/23/433228503/in-spain-entire-villages-are-up-for-sale-and-theyre
-going-cheap
9
[Link]
7.3 Technological Change and Disruption 199

that new technology is never adopted by 100% of the population and that older
technology survives in niches around the world.

⇨ Exercise 7.3
Find an example of an old technology that should in theory be “obsolete” but
that is still in active use today. First describe the “old” technology and how
you found it. Then describe the “new” technology or technologies that
replaced it and attempt to explain the reasons (preferably using both qualita-
tive and quantitative arguments) why the users of the old technology never
adopted newer alternatives, or why an old technology was reborn.

7.3 Technological Change and Disruption

Rogers (1962) focused almost exclusively on the adoption of a single new solution
or technology and its rates and patterns of diffusion in society. He also described in
detail the characteristics of different types of adopters. However, he did not go all
the way to considering multiple waves of technological change (technology B
replaces A, and B subsequently gets superseded by C, etc.). History shows us that
once a technology has been widely adopted, it may be “toppled” or superseded by a
newer one, and this may happen multiple times.
Jim Utterback (1994) of MIT in particular has studied such waves of technologi-
cal innovation and how older technologies are replaced with newer ones over time.
In his book titled “Mastering the Dynamics of Innovation,” he studies not only dif-
ferent waves of technology but also their impact on the underlying industrial struc-
ture. Fig. 7.12 shows some of Utterback’s case studies.
At first, the dynamic is quite simple to understand (see Fig. 7.13). An established
technology or product is adopted and broadly diffused. With improvements in per-
formance and lower cost, the market for the product (or service) is expanded and
new competitors jump onboard since the growth attracts new competitors. An
incumbent industrial base is established and gradually improves the product over
time through a combination of product and production process improvements. In
parallel innovators are working on a new technology, with the same underlying
function (see Chap. 1) such as “document writing,” “food preserving,” “light pro-
viding,” “glass producing,” and “image capturing,” referring back to the examples
in Fig. 7.12.
At some point in Fig. 7.13, the incumbent product’s rate of improvement slows
considerably (t1) due to diminishing returns. The new technology is “pushed” by its
creators and embodied in a new “invading” product or service. However, initially
the new product is inferior, since it is less perfected. Over time, however, its perfor-
mance improves faster and faster and eventually matches (t2) and then surpasses the
incumbent technology, gradually displacing it.
Reality may be more complex in that the owners of the established product (or
service) may see the leading edge of substitution and may “fight back” by renewing
200 7 Technological Diffusion and Disruption

Fig. 7.12 Waves of innovation and change studied by Utterback (1994)

Fig. 7.13 Performance of

an established and an
invading technologically
enabled product. The burst
of improvement in the
established product occurs
in response to the invader.
(Source: Utterback 1994)
7.3 Technological Change and Disruption 201

their efforts and, thus, producing a “burst of improvement” in the established prod-
uct. This, however, only delays the inevitable, which is that the new technology –
which is usually based on a different system architecture and physical working
principles – takes over all or a large majority of the market share after (t3). This
dynamic is conceptually rendered in Fig. 7.13.
If the displacement of the established product (and associated technology) is due
to a set of new players, causing the decline and even bankruptcy of the established
players, we call this phenomenon a technologically-induced disruption, or simply
“disruption.” The enabling technology underpinning the successfully invading prod-
uct is termed a “disruptive technology.” Note, however, that Christensen (below) has
introduced a somewhat different definition for what is meant by “disruption.”

➽ Discussion
Can you cite an example of an invading product technology that displaced an
established technology? When and why did that happen?

As we will see in the next section, rarely are firms able to disrupt themselves, and
the Innovator’s Dilemma claims to explain why. Those firms that are not able to
switch at the right time from the old to the new technology will most likely cease
to exist.
It is interesting to consider some historical cases of technological disruption.
Utterback (1994) goes into considerable depth in the case of the ice-harvesting
industry (for refrigeration of meat, dairy, drinks, hospitals, etc.) in the United States
in the nineteenth and early twentieth centuries. The ice-harvesting industry was
centered in New England, and one of its pioneers was Frederic Tudor of Boston, the
so-called Ice King. His likeness is shown in Fig. 7.14 (right), and the initial method
of ice harvesting from frozen ponds was crude and labor intensive (for the laborers
and their horses), as shown in Fig. 7.14 (left).

Fig. 7.14 (left) Ice harvesting on Spy Pond 1854, Arlington MA, USA; (right) Frederic Tudor the
“Ice King” (1783–1864)
202 7 Technological Diffusion and Disruption

Tudor built an ice distribution empire that served the Southern United States, the
West Indies, Europe, and even India. His company thrived for decades but was dis-
rupted by mechanical ice-making machines in the late nineteenth century.
As demand for ice grew with an expanding U.S. population and demand overseas
(Tudor shipped to places as far away as New Orleans, San Francisco, the Caribbean
Islands, Cuba, Brazil, India, and even Hong Kong), there was a need to increase the
production rate and reduce cost. A key technological improvement was the inven-
tion of the “ice plow,” which was patented by Nathaniel Jarvis Wyeth in 1825. The
ice plow was a cutting device shown in Fig. 7.14 (left), which harnessed the power
of horses in a way that led to uniformly shaped blocks of ice which both reduced
cost (by about a third) and improved the quality of the ice product, including its
transportation. To minimize the loss of ice during transport, especially in warmer
climates, a whole supply chain was set up with “ice houses” at major ports (e.g., in
Havana in 1816) and optimized insulation and stacking, which included extensive
use of sawdust, which was also readily available in New England as a by-product of
the timber industry. The resulting expansion of the New England ice industry in the
nineteenth century was impressive (see Fig. 7.15).
One of the major issues for customers of ice in the South and away from major
ports was the seasonality of prices and availability of ice. While the average price of
a ton of ice, for example, in Charleston, South Carolina, dropped from $166 in 1817
to $25 in 1834, it was the volatility of ice prices (e.g., between $6-8/ton in a “good”
year to $60-$75/ton in a poor year due to a mild winter and diminished ice produc-
tion) that caused problems for customers. The response to this was the development
of mechanical means of ice production using the Carnot process (see Fig. 1.1). This
required a compressor, refrigerant working fluid, condenser, evaporator, and heat
exchangers, among others. While the first “artificial” ice was made in 1755, it took
until the 1850s until viable ice-making plants could exist. Initially, ice machines
were inferior to natural ice, however, over time due to experimentation with differ-
ent refrigerants (e.g., Boyle’s ammonia compression machine in 1872 became an

Fig. 7.15 Quantity of ice shipments from New England between 1806 and 1856. (Source: Based
on data in Henry Hall, The Ice Industry of the United Sates with a Brief Sketch of Its History and
Estimates of Production, U.S. Department of the Interior, Census Division, Tenth Census, 1880, v.
22 (Washington, D.C.: U.S. Government Printing Office, 1888, reprinted by the Early American
Industries Association), p. 3)
7.3 Technological Change and Disruption 203

Fig. 7.16 Ice-making

plants in the United States
between 1869 and 1920.
(Source: U.S. Bureau of
the Census, cited in
Cummings, The American
Ice Harvests, p. 11, and
Jones, America’s Ice Men,
p. 159)

enabling technology) and other improvements, they became viable. In 1868, New
Orleans opened its first local ice-making plant that produced at $35/ton. In 1889,
there were 222 U.S. ice-making plants in operation. The innovators were in the
South, not in the North, where the new technology (mechanical ice making) was the
most competitive. Fig. 7.16 shows the exponential growth in the number of ice-­
making plants in the United States between 1869 and 1920.
Initially, the emergence of the new technology did not have an immediate effect
on the New England ice producers as the underlying market was rapidly expanding
along with the burgeoning U.S. population.10 After a few years, however, the threat
became apparent and the incumbent ice harvesters did not lay down passively. They
redoubled their efforts and made large investments and technological improvements
such as steam-powered circular saws for ice cutting, insulated railroad cars, improved
ice houses, and what today we would call the predecessor of a “cold chain” that could
extend for thousands of miles. This worked for a while (see the “burst” of improve-
ment in Fig. 7.13) and natural ice shipments peaked in 1886 at 25 million tons.
The machine-made ice, however, kept improving relentlessly both in terms of
quantity and quality of production. While initial production costs were initially any-
where from $20 to $250 per ton, the aim of the inventors was to make ice for a cost
as low as $0.75–$1.00 per ton.11 The higher transportation costs from the North
could not compete with this and as a result the natural ice-harvesting industry even-
tually collapsed in the early twentieth century. This was epitomized when in 1909
Massachusetts – the last bastion of natural ice harvesting – opened its first

10
Utterback (1994) notes that a rapidly growing underlying market can “mask” an ongoing disrup-
tion because absolute sales numbers of the incumbent technology can continue to grow, even as the
relative market share of the incumbent technology drops. This is especially true during a historical
period where sales numbers, quarterly reporting, and industry-wide market surveys were scarce or
wholly unknown.
11
This is the first instance where we mention the concept of figure of merit (FOM)-based target
setting for technologies. The artificial ice machine makers set themselves a target of $1/ton of ice
produced in the 1880s, which was roughly a 20-fold improvement of what was possible in the late
1860s. This concept of technology target setting will feature prominently in Chap. 8 on Technology
Roadmapping.
204 7 Technological Diffusion and Disruption

Fig. 7.17 US ice exports

between 1850 and 1910.
(Source: U.S. Census
Bureau, cited in
Cummings, The American
Ice Harvests)

mechanical ice-making plant. The decline in natural ice harvesting is illustrated in

Fig. 7.17, which shows the initial rise and then the decline of US ice exports
over time.
The mechanical ice-making industry itself was later disrupted in a major way
after World War I (1914–1918) when electrification of the country as depicted in
Fig. 7.10 allowed for electro-mechanical refrigerators (see Fig. 1.1) in homes,
which were almost completely adopted between 1920 and 1950. This obviated the
need for major ice production, even though it still exists as a niche market today. As
Fig. 7.12 suggests, the development of aseptic packaging of food may be a further
disruption in this area, understanding that the ultimate major function for which the
ice was used is to prevent food from spoiling.
This case and the associated “disruption” can be better understood if the charac-
teristics of alternative and, therefore, competing technologies can be expressed
using the same relevant figures of merit (FOM) that capture “value” to the users or
adopters such as [$/ton] of ice.

7.4 The Innovator’s Dilemma

The concept and framework of the “Innovator’s Dilemma” were first proposed by
Clayton Christensen in 1997. His book has attracted a large following in manage-
ment and academic circles and among entrepreneurs and helped clarify the notion
of so-called “disruptive technologies.” An important point that Christensen makes
is that disruptive technologies are not those that lead to gradual incremental or even
radical (step-wise) improvements in an existing technology or product – those are
referred to as “sustaining” innovations – but a technology that has the potential to
displace and destroy entire incumbent firms and industries.
Examples of technologies that have – in hindsight – proven to be disruptive are
shown in Table 7.1. This is by no means an exhaustive list.
This phenomenon has claimed many victims according to Christensen. Iconic
and leading firms such as Kodak (photography), Digital Equipment Corporation
7.4 The Innovator’s Dilemma 205

Table 7.1 Comparison of established versus disruptive technologies (examples)

Established technology Disruptive technology
Silver halide photographic film Digital photography
Ice for refrigeration Electro-mechanical refrigerator
(natural or from machine)
Horse Automobile
Offset printing Digital printing
Crewed fighter aircraft Unmanned aerial vehicles (UAVs)
Open surgery Laparoscopic and endoscopic surgery

DEC (computers), or Sears (retail) no longer exist or are a shadow of their former
selves because they were disrupted by new technologies. New technologies that, in
some cases, they themselves initiated, and new technologies that in some cases were
coupled to new business models.
The essence of the Innovator’s Dilemma is that the very best practices that have
been traditionally taught in many business and engineering schools: continuously
improving products, listening carefully to customers, moving into higher perfor-
mance categories that have the potential for higher margins (profits), etc. are exactly
the reasons that have caused the decline in incumbent firms due to a failure to rec-
ognize the transformative potential of disruptive technologies. Sometimes it is bet-
ter not to listen to (existing) customers, to pursue smaller or even nonexisting
markets, and to launch products (or services) with – at least initially – lower margins
than those in existing large markets.
The principles of disruptive innovation are:
1. Recognizing the difference between sustaining and disruptive technologies.
Sustaining technologies improve existing products (or services) for well-­
established figures of merit (FOMs) over time. Even a large improvement in an
existing FOM (as opposed to a small incremental step) is not disruptive. It can
instead be termed as a radical sustaining innovation. A good example is the
replacement of piston engines with turbojet engines in civil aviation. It was a
radical innovation (increasing the speed of aircraft significantly), but it did not
fundamentally change the nature of the market, the leading firms, or industry
structure. Disruptive innovations often initially yield inferior performance on an
established FOM, while offering something new of value on a different FOM, but
usually to a different group of customers. They offer a different value proposition.
2. Differential rate of progress between technology and market demand. In
many instances, uncovered by Christensen, the annual rate of progress achiev-
able or achieved by a technology in terms of dFOM/dt (see Chap. 4) exceeds that
which is demanded by the market or what the market is willing to pay for. Once
performance is “good enough” for a given market segment, the producing firm
will then typically seek higher-end applications and markets, which may have
higher margins (per unit), and can make use of the higher-end performance. This
“overshoot,” however, creates a potential opening for a disruptive competitor
from below. This situation is shown in Fig. 7.18.

Licensed to ([Link]@[Link])
206 7 Technological Diffusion and Disruption

Fig. 7.18 The impact of sustaining and disruptive technological change according to Christensen.
The sustaining technologies increase product performance over time in existing markets. According
to this model, disruptive technologies “invade” from below with initially a lower level of perfor-
mance but other advantages, enough to grow and displace the incumbent technology over time

3. Disruptive technologies versus “rational” R&D investments: Established

firms are mostly driven by an internal dynamic that favors investing in sustaining
technologies to support their established “cash cow” products that provide high
margins. Investing substantially in disruptive technologies is not viewed as a
rational decision by the management of established firms since disruptive
­technologies initially look inferior, less profitable, and more uncertain. However,
a “wait and see” attitude is rarely rewarded in practice, since by the time the new
market has fully developed, the “disruptors” are significantly ahead in market
share and a full understanding of the new value chains.
Similar to Rogers (agriculture) and Utterback (ice production), Christensen
selected one major case study in depth to develop his analysis and framework.
Christensen did an in-depth analysis of the market and underlying technologies
for computer disk drives for data storage. In the early days of the emergence of
mainframe computers, there was a need for increasing amounts of data storage. In
order to maintain the same physical size of disk drive (14-inch was the initially
dominant size, followed by 8-inch), manufacturers invested heavily in R&D to
increase recording density. Fig. 7.19 illustrates the sustaining character of sequen-
tial technologies such as ferrite-oxide heads, thin-film heads, and magneto-resistive
heads to increase the data capacity per disk. These technological innovations
achieved three orders of magnitude improvement between 1975 and 1995 and had a
clearly sustaining character.
7.4 The Innovator’s Dilemma 207

Fig. 7.19 Impact of new read-write head technologies in sustaining the trajectory of improvement
in recording density for computer storage disks (Source: Christensen; Data are from various issues
of Disk/Trend Report)

Fig. 7.20 A disruptive technology change: The 5.25-inch Winchester Disk Drive (1981)(Source:
Data are from various issues of Disk/Trend Report)

In about 1981, a new disruptive technology, the Winchester-type 5.25-inch drive,

was developed and proposed as an alternative to the 8-inch drive for minicomputer
manufacturers. The specifications of the 8-inch and 5.25-inch drives at the time are
compared in Fig. 7.20.
Since the dominant figure of merit for minicomputer manufacturers was the cost
of a megabyte of storage [$/MB], the 5.25-inch drive was clearly not competitive
208 7 Technological Diffusion and Disruption

and was not taken up by minicomputer manufacturers at that time. However, two
other FOMs, namely the physical volume (smaller by a factor of ~4) and unit cost
(33% cheaper) were attractive to the just emerging developers of smaller desktop
computers (which were not believed to be a serious market threat by incumbent
manufacturers of the larger minicomputers). However, as the 5.25-inch drives were
being adopted by the new desktop market, they grew in capacity by an estimated
50% per year and eventually intersected the actual requirement of the minicomputer
market by about 1985. As Christensen states:

*Quote
“As in the 8-inch for 14-inch substitution, the first firms to produce 5.25-inch
disk drives were entrants; on average, established firms lagged behind entrants
by 2 years. By 1985, only half of the firms producing 8-inch drives had intro-
duced 5.25-inch models. The other half never did.”

This pattern repeated itself again for the smaller 3.5-inch and the-2.5 inch drives
(Fig. 7.21). Each time, a significant number of disk drive manufacturers who only
focused on one larger market and the one dominant FOM in that market had diffi-
culty in recognizing the potential of the smaller – initially inferior – disruptive tech-
nology. This was true even though the established firms had very capable
management and R&D departments. However, the tendency to stay entrenched in
the existing market and investing the vast majority of resources (people and money)
in R&D for the sustaining technologies was so dominant that new ideas and con-
cepts (such as the smaller more compact drives that had advantages other than what
the marketing department found existing customers wanted) were killed off early or
recognized too late.
One of the important concepts here is the idea that technology does not exist in
isolation (whether sustaining or disruptive) but is embedded in value networks of
nested supply chains of component providers, subsystem integrators, and original
equipment manufacturers (OEMs). This is shown in Fig. 7.22, and as can be seen,
the magnetic disc drive technology is embedded in a larger industrial ecosystem that
forms clear rules of competition and expectations over time.
The example shown in Fig. 7.22 is for a 1980s vintage management information
system (MIS) enabled by a mainframe computer.
One of the most important characteristics of such a value network is that it is
centered around one or several key enabling technologies and that at the boundaries,
the rank order of priority in terms of FOMs is clearly defined (e.g., for disk drives
in mainframe computers, the total data storage capacity per dollar is essential while
the physical volume is ranked much lower). As Christensen states it: “The way
value is measured differs across networks. In fact, the unique rank-ordering of the
importance of various product performance attributes defines, in part, the boundar-
ies of the value network.”
This can then explain why a disk drive such as the 5.25 inch was initially unat-
tractive to the minicomputer market, since it was inferior in many of the product
7.4 The Innovator’s Dilemma 209

Fig. 7.21 Intersecting trajectories of capacity demanded versus capacity supplied in rigid disk
drives. For example, disruption of 8-inch drives (B) by 5.25-inch drives (C) in the minicomputer
market occurred in 1987 when the smaller drives met the performance demands of the larger mar-
ket, at a lower cost, and for less volume. The 8-inch drives were displaced as a result and had to
retrench to the higher-end but smaller market for mainframes. (Source: Clayton M. Christensen,
“The Rigid Disk Industry: A History of Commercial and Technological Turbulence.” Business
History Review 67, no. 4 (Winter 1993): 559. Reprinted by permission)

performance attributes (which we call FOMs in this book) that mattered the most to
the customers of that market such as capacity [MB], cost per unit of data stored [$/
MB], and access time [ms]. In parallel, a different value and industry structure
emerged for the desktop computer (and much later laptops and mobile devices such
as tablets) which rank-ordered cost per unit, volume, and weight much higher.
210 7 Technological Diffusion and Disruption

Fig. 7.22 A nested system of product architectures defining a value network

Once 5.25-inch disk drives were adopted in the lower-end market and that mar-
ket developed, they continued to improve at a very rapid rate (40–50% per year in
terms of storage capacity), eventually catching up to the requirements of the higher-­
end market that had initially shunned the 5.25-inch technology. Now, however, the
5.25-inch drives did not only meet the capacity requirements of minicomputers in
terms of storage capacity but also brought with them all the other advantages that
they had inherited from the lower-end market (such as lower weight, volume, power
consumption, and vibrations), eventually disrupting the 8-inch disk market
completely.
Besides a clearer definition of what is meant by “disruptive technology,”
Christensen also highlights the fact that, as products become commoditized over
time, the key FOMs that drive competition in the market change in discrete phases,
see Fig. 7.23.
During the initial phase 1, the competition in computer disk drives was driven
primarily by capacity (who can provide the most data storage in megabytes?).
Eventually, as the market needs were largely satisfied in terms of capacity, competi-
tion shifted to physical size in phase 2. Once further reductions in the size of a
computer were no longer seen as valuable (i.e., the shadow price of a cubic inch of
computer volume approached zero), the focus shifted to reliability in phase 3 and
finally, price, in phase 4. Each time a switch in the rank order of FOMs occurred,
there was a discontinuity in the market enabled by a disruptive technology.
In the next chapter (Chap. 8), we will focus on the topic of Technology
Roadmapping, which dwells precisely on the issue of which technologies are needed
(both sustaining and disruptive) to enable a firm to be competitive over time in both
7.5 Summary 211

Fig. 7.23 Changes in the basis of competition in the disk drive industry

existing and newly emerging markets. This can be a significant challenge if a firm is
engaged in multiple different markets and value chains (such as the one in Fig. 7.22)
at the same time. The key issue is to set realistic FOM targets and to derive from
them an appropriate allocation of technology and product development projects in
their R&D portfolio.

7.5 Summary

This chapter gave an overview of important concepts in technology adoption and

diffusion (which generally follow an S-Curve-type behavior). This includes the fact
that some individuals and organizations may be innovators and early adopters while
others are late adopters depending on the level of information and risk tolerance
they have when it comes to the prospects of new technologies, or new products
enabled by them.
The disruption of existing markets or value chains can occur when the rank order
of figures of merit (FOM) is switched and customers are basically satisfied with the
level of performance of an existing technology but start weighting other FOMs
more heavily. This was clear in the computer disk drive industry (Christensen 1997),
where storage capacity was dominant in early mainframes, until size, reliability, and
eventually price became dominant considerations in the emerging desktop com-
puter industry.
We can see this switch of priorities happening in aviation as well (see case study
2 in Chap. 9) where initial competition was driven primarily by range and payload,
212 7 Technological Diffusion and Disruption

followed by reliability and safety, fuel burn as a proxy for operating costs, and now
increasingly environmental impact (CO2-equivalent emissions).
In this light, we can argue that the switch from naturally-cut to machine-made ice
in the refrigeration industry was not really a “disruptive” innovation, but rather a
radical-sustaining technological innovation, since the dominant FOM was still [$/
ton] of ice and ice was still used for cooling. The same railroad cars and ice houses
that were used for naturally harvested ice could also be used for machine-made ice.
However, the switch from using ice to domestic refrigerators powered by electricity
was disruptive (in the sense of Christensen) since it began with much smaller units
(the household) and, in a distributed way, eventually obviating the need for trans-
porting ice over large distances.

⇨ Exercise 7.4
Select an example of a disruptive technological innovation, describe it in
some detail, and argue why it should not be considered either as an incremental-­
sustaining or radical-sustaining innovation. If possible provide some quantita-
tive numbers over time and show the rank order of key FOMs.

Appendix

 atlab Code for Agent-Based Simulation

M
of Technology Diffusion

% simple Rogers-type agent-based diffusion model

N=10000; % population size
a=zeros(N,1); % initially all are non-adopters
T=100; % time periods
C=1; % number of contacts per time period
n=0.5+(1/6)*randn(N,1);% create Gaussian distribution of population
[fa,ind]=min(n); a(ind)=1; % find and set initial adopter
for t=1:T; % loop over all time periods
inda=find(a); % find all adopters at this time step
A(t)=length(inda); % record the number of adopters at time t
for inda2=1:length(inda); % loop over all current adopters
indr=randi(N,C,1); % create random contacts this time period
for inda3=1:length(indr); % loop over all random contacts
nr=rand(1); % generate uniform random decision variable
if n(indr(inda3))<=nr && a(indr(inda3))==0; % test adopt
a(indr(inda3))=1; % nr adopts
end
end
end
end
References 213

References

Christensen, Clayton M., “The Innovator’s Dilemma – When New Technologies Cause Great Firms
to Fail”, Harvard Business Review Press, 1997, ISBN: 978-1-63369-178-0
David PA. Clio and the Economics of QWERTY. The American Economic Review. 1985 May
1;75(2):332–337.
Doufene, A., Siddiqi, A., & de Weck, O. (2019). Dynamics of technological change: nuclear
energy and electric vehicles in France. International Journal of Innovation and Sustainable
Development, 13(2), 154–180.
Griliches, Z. (1957). Hybrid corn: an exploration in the economics of technological change.
Econometrica, 25(4), 501–522.
Liebowitz SJ, Margolis SE. The fable of the keys. The Journal of Law and Economics. 1990 Apr
1;33(1):1–25.
Rogers, Everett M., “Diffusion of Innovations”, First Edition, The Free Press, A Division of Simon
& Schusters Inc., 1962, Fifth Edition, 2003, ISBN-13: 978-0-7432-2209-9
Utterback, James M., “Mastering the Dynamics of Innovation”, Harvard Business School Press,
Boston, Massachusetts, 1994, ISBN 0-87584-342-5
Chapter 8
Technology Roadmapping

Advanced Technology Roadmap Architecture (ATRA)

Inputs
Steps Outputs
Strategic Drivers for Technology 8
+10y Technology
1. Where are we today? FOMjj Roadmaps
L1 Products and Missions +5y

L2 Technologies Technology State of the Art and Organization Figures of Merit (FOM)
Competitive Benchmarking Current State of the Art (SOA)
Competitor 1 Technology Trends dFOM/dt
Technology Systems Modeling Competitor 2

Tech Pul
Pull Today FOMi
2. Where could we go?
Dependency Structure Matrix

L1
+10y
Scenario-based
Technology Systems Modeling and Scenario A Technology Valuation
FOMj

Trends over Time

Technology +5y Design Reference Missions
Scenario B
Projects Future Scenarios
? T h l
Technology V
Valuation
l ti
3. Where should we go? Vector Charts
L2
Scenario Analysis and FOMi
Technology Valuation
Tech Push Technology Investment
Efficient Frontier
E[NPV] - Return

Technology Scouting 4. Where we are going! Pareto-optimal set of technology

Knowledge Management Technology investment portfolios
Technology Portfolio Valuation, Portfolio Recommended Technology Portfolio
Intellectual Property Analytics Optimization and Selection Technology (Expected NPV and Risk)
Projects
σ[NPV] - Risk
Foundations C
Cases

Definitions History Nature Ecosystems The Future Case 1 Case 2 Case 3 Case 4
What is Milestones of Technology, Nature Technology Diffusion, Is there a Deep Space DNA
Singularity ? Automobiles Aircraft
Technology? Technology and Humans Infusion and Industry Network Sequencing

© Springer Nature Switzerland AG 2022 215

O. L. de Weck, Technology Roadmapping and Development,
[Link]
216 8 Technology Roadmapping

8.1 What Is a Technology Roadmap?

A technology roadmap is a plan that shows which technologies will be used by

which current or future product (or service or mission) and by when these technolo-
gies have to be ready and at what level of performance. There are different levels of
sophistication and detail when it comes to technology roadmaps. Some are very
simple (essentially a single slide or page) and show a map over time, as in Fig. 8.1,
while some others are more advanced, containing links to quantitative models,
trends over time, and benchmarking against the competition.
The history of technology roadmapping goes back about 50 years but became
more prominent in 1987 when Motorola published its “Motorola’s technology road-
map process” (Kerr and Phaal 2020). Since the 1990s, academic research on tech-
nology roadmapping has increased and technology roadmaps are often also
mandated by government entities, such as the US Congress, the United Nations, and
other organizations. The main point of technology roadmapping is to give stake-
holders a sense of direction and to ensure that R&D resources are deployed in a
targeted manner.
The purposes of technology roadmaps in an organization are to:
• Show the relationships across technologies, capabilities, products/services,
and needs.
• Align investments in technology and in the new development of new capabilities
to deliver on future market needs.
• Map technologies to products/missions/services and define a timeline for matu-
ration and technology adoption.
More advanced technology roadmaps contain models of the needed technologies
across multiple levels of decomposition and have clear links to figure of merit
(FOM) targets (see Chap. 4) and ongoing and future internal and external R&D
projects to achieve these targets.
There is a large diversity and difference between the state of the art, that is, what
research tells us and what the most advanced companies do in roadmapping, and the
state of practice, that is, what most average companies do. Generally, technology
planning and roadmapping is the responsibility of the Chief Technology Officer
(CTO) of the company, occasionally that of the Head of Engineering, typically at
the Vice President, or Senior Vice President level.1
According to Bernal et al. (2009), there are different purposes and flavors of
technology roadmaps. Some of these are listed below, even though the most basic
one is shown in Fig. 8.1:
• Product planning.
• Capability development.

1
I served as Senior Vice President (SVP) of Technology Planning and Roadmapping at Airbus for
2 years (2017 and 2018) while on leave from MIT and reported to the Chief Technology Officer
(CTO). The CTO at Airbus is at the Executive Vice President (EVP) level and is a member of the
company’s senior executive management team (the so-called “C-Suite”).

Licensed to ([Link]@[Link])
8.1 What Is a Technology Roadmap? 217

Fig. 8.1 “Simple”

technology roadmap:
linking technologies (blue
at the bottom) against the
products that will
implement them (red at the
top) along a timeline.
Interdependencies exist
between products and
technologies. (Source:
Bernal et al. 2009)

• Strategic planning.
• Long-range planning.
• Knowledge planning.
• Project planning.
• Integration planning.
Figure 8.2 shows a different type of technology roadmap, focused on capabilities
instead of products. Capabilities are functions or processes and “know how” that an
organization acquires over time to create new products and services (or improve
existing ones). An example of that could be the capability to send data through
space using light, as opposed to radio waves. Both of them are in the radio-­frequency
spectrum, but deep-space optical communications (see Chap. 13: DSN case study)
require very different mathematics, physics, equipment (telescopes vs. antennas,
lasers vs. masers), software, and operating procedures.
The top line in Fig. 8.2 shows “Events” which could be internal or more often
external events that act as pacesetters for market and business tendencies and trends.
An event could be a future planned space mission, or it could be a major trade show
at which a new product or service will be introduced. These then act as triggers for
capability development, which in turn provides a “pull” for new or improved tech-
nology development. As in Fig. 8.1, the x-axis represents time as it is very important
that the pacing (speed) of technology development is clearly defined and linked to
the external trends, triggers, and events.
An important point that is often missed in technology roadmapping is that not all
technologies are created equal. Regardless of which physical, chemical, or biologi-
cal working principle a technology relies on, it has different roles to play in future
products, services, or capabilities.
Table 8.1 shows a distinction between sustaining and disruptive technologies.
We have already encountered this in Chap. 7 when discussing the innovator’s
dilemma. Sustaining technologies are those that provide improvements to existing
products along well-known and accepted figures of merit (FOM). If the progress is
small we speak of incremental improvement, and if the progress is significant or
rapid we speak of radical-sustaining improvement. Disruptive technologies are
those that provide something new or different along a different FOM than what is
currently valued by the established market.
218 8 Technology Roadmapping

Fig. 8.2 Technologies map to capabilities that map to markets and events

Table 8.1 Distinction between different technologies by role

Technology type Role of technology
Sustaining-incremental A technology that is aimed at enhancing an existing product or
(small improvements) service by making a small but positive change, typically on the order
of 1–5% improvement in a known FOM
Sustaining-radical A technology that is aimed at enhancing an existing product or
(large improvements) service by making a large positive change, typically on the order of
>5% improvement in a known FOM
Disruptive A technology that significantly shifts the competition to a new regime
where it provides a large improvement on a different FOM than the
mainstream product or service (see Chap. 7)
Enabling A technology that is absolutely required (at a given level of
performance) since without it the product or mission cannot happen
Supporting A technology that is contributing to a new product or mission directly
or indirectly, but whose availability is not absolutely required. In
some cases, there may be alternatives available

Another way to distinguish technologies is between supporting and enabling

ones. Supporting technologies make a product or mission better, but if the technol-
ogy was not present, the product or mission could still happen. Enabling technolo-
gies on the other hand are sine qua non.
Another, perhaps more organizationally oriented form of technology roadmap is
the knowledge roadmap, depicted in Fig. 8.3. It is driven by business goals and
planned projects and activities, knowledge and enablers, and processes related to
knowledge as well as intellectual resources that are needed to succeed. This includes
8.1 What Is a Technology Roadmap? 219

Fig. 8.3 Knowledge planning: Aligning intellectual resources, capabilities, processes, projects,
and business objectives. We will discuss knowledge management in Chap. 15

experts, databases, procedures, software, and training courses required, among oth-
ers. Say, for example, an automotive manufacturer decides to switch from internal
combustion engines to electric drives (see Chap. 6); this will require the establish-
ment of new competencies and knowledge in the firm, for example, for developing
and testing of high voltage motors, switches, power conditioning equipment, and
batteries.
As stated earlier, technology roadmapping has been practiced informally in
industry since the 1960s, and scholarship on roadmapping has blossomed, roughly
since the mid-1990s. One individual who has fully dedicated himself to the aca-
demic study and industrial application of technology roadmapping is Dr. Robert
Phaal at the University of Cambridge (UK) (Kerr and Phaal 2020).
Figure 8.4 shows the roadmapping framework proposed by Phaal and Muller
(2009), and it captures a kind of integrated “metaview” of the different flavors of
roadmaps shown in Figs. 8.1, 8.2 and 8.3. The Cambridge framework for technol-
ogy roadmapping has the following features (from left to right):
• Roadmapping is considered from different viewpoints (commercial and strate-
gic; design, development, and production; and technology research).
• The technology roadmap has an architecture, meaning a logical structure with
elements that clearly relate to each other through perspectives: market, business,
product, service, system, technology, science, and resources.
• The roadmap framework elicits these different elements and links them across
the timeline including past, short-term (typically 1–3 years), medium-term
(3–10 years), and long-term (>10 years), as well as a long-range vision. The
result of applying this framework should be a strategic and aligned plan for pur-
poseful innovation in the organization.
220 8 Technology Roadmapping

Fig. 8.4 A potential technology roadmapping framework. (Phaal and Muller 2009)

• As a result of the technology roadmapping framework, there are different ques-

tions answered: When something is needed? What is needed? Why is it needed?
• The type of information and knowledge contained in a roadmap includes strate-
gic drivers, market needs, form, function, and performance levels of future
­products and services, as well as the solutions and resources needed to imple-
ment the plan.
In a later section, we will see a different – and yet related – implementation of
technology planning and roadmapping that embodies the “MIT approach” to tech-
nology roadmapping. There are a few important points to keep in mind for organiza-
tions that decide to adopt technology roadmapping. Some of these are often
overlooked, which can substantially reduce the benefits obtained from roadmapping:
• Technology roadmapping is not a purely technical function, it requires bringing
together people from marketing, strategy, engineering, research, manufacturing,
procurement, finance, and even HR under a common umbrella. It is one of the
most multidisciplinary activities that can be done in a technology-enabled firm.
• Creating a set of high-quality technology roadmaps is not a side activity but cen-
tral to the long-term success and survival of the company. As we saw in Chap. 7,
the lack of foresight and proper balance between sustaining and disruptive inno-
vations can lead to the downfall of an enterprise. This can happen surprisingly
quickly.
• The effort for creating and maintaining technology roadmaps is substantial if it
is to be done well with real impact on the direction of the R&D portfolio. While
it is useful to begin with a set of qualitative workshops, eventually the roadmaps
should be maintained and refreshed regularly by dedicated roadmap owners
(RMOs).2

2
We estimate that it takes about $250 K per year (2019 figures) to create and properly maintain a
quality technology roadmap. This means that an organization that has about 20 technology road-
8.1 What Is a Technology Roadmap? 221

Fig. 8.5 Example roadmap structure (“architecture”) proposed by Phaal and Muller (2009)

• It is important that technology roadmaps are well organized and somewhat stan-
dardized such that different technologies can be compared on an equal footing.
Figure 8.5 presents a roadmap structure as proposed by (Phaal and Muller 2009).
At the top of the roadmap is the market and business view. In what markets and
segments is the company active today? Where does it want to compete in 3 years?
In 10 years? What are the different business units (BUs) and what is their competi-
tive position?
What are the different products and services offered by the company? The exam-
ple provided here is from a European Tier 1 supplier of off-highway vehicles, there-
fore the list of their products and services contains things such as: wheels, axles,
transmissions, driveline systems, tractor attachments and hitches, cabs, as well as
product distribution and servicing.
In terms of technologies, the firm has identified the following main technologies
it perceives as enabling or enhancing (see Table 8.1): computer-aided engineering
(CAE), manufacturing (e.g., milling and casting), electronics, driveline technolo-
gies, materials, and other.
Finally, the roadmap lists at the bottom the resources needed to actually imple-
ment the roadmap in practice. This includes finance (what are the necessary R&D
project investments?), skills and competencies (impacts on HR planning, recruiting,
and training programs), alliances, and supply chain impact (are we doing everything

maps should plan to spend about $5 million per year on technology roadmapping.
222 8 Technology Roadmapping

alone? Are we partnering? Make or buy?). And finally any impacts or actions to be
taken with respect to the firm’s organization and culture.
From our experience in creating and implementing technology roadmaps at sev-
eral global Fortune 500 firms, there are a few lessons learned:
• Technology planning and roadmapping should have the full support of the CEO,
CTO, Head of Engineering, and the board. Without that active support, it becomes
a less than impactful activity.
• Technology roadmaps must be validated with quantitative technical and financial
models. Many technology roadmaps in practice are purely qualitative in nature.
It is, therefore, “easy” to make plans and set quantitative FOM targets and so
forth. However, without a quantitative analysis, whether these targets are (i) too
easy, (ii) about right, or (iii) too difficult to achieve within the resources and
timeframe available, technology roadmaps will not have much credibility. In
other words, technology roadmaps need to be validated by data, analysis, and an
organized review process involving experts and senior management.
• Individuals who are selected for technology roadmapping should be a mix of
personnel from more experienced technical staff (e.g., chief engineers, senior
technical experts, and chief scientists) and more junior staff such as new research
scientists and junior engineers. The more senior staff will typically focus more
on sustaining incremental innovations and throw up warning flags why some-
thing new cannot or should not be done. The junior staff will typically push for
more radical and disruptive innovation. This dialogue and tension is healthy and
can lead to a well-balanced strategy.
Next, we will consider an example of a “complete” roadmap for a new product
in an aerospace company based on solar-electric flight. This roadmap is based on
publicly available information.

8.2  xample of Technology Roadmap:

E
Solar-Electric Aircraft

August 10, 2018, was an exciting day for aviation. An aircraft named “Zephyr”
made history and established a new world record for sustained flight of a heavier-­
than-­air aircraft without burning a drop of fuel.
The aircraft is a solar-electric unmanned aerial vehicle (UAV) flying at the edge
of the stratosphere and at an altitude of about 70′000 feet, twice as high as most
commercial airliners. See Fig. 8.6 for an infographic on Zephyr which is designed
and manufactured by Airbus Defense and Space, and was originated by the firm
QinetiQ (2003) based on an earlier project at Newcastle University in the UK.
While solar-electric aircraft have been developed for the last three decades or so,
it is only now that the enabling technologies, such as thin-film photovoltaics (see
Fig. 4.12), lithium-based rechargeable batteries, lightweight composite structures,
and miniaturized electronics (payload cameras and communications electronics),
8.2 Example of Technology Roadmap: Solar-Electric Aircraft 223

Fig. 8.6 Zephyr solar-electric aircraft infographic (World Record 2018)

have progressed to the point where sustained flight based only on solar energy
through the day-night cycle has become possible. The endurance world record that
Zephyr established in Arizona in 2018 stands at 25 days, 23 hours, and 57 min-
utes. This record is sure to be broken in the coming years, but what will it take?
In this section, we provide a notional technology roadmap for solar-electric air-
craft as a new business category. The potential market and business applications for
this type of aircraft, also known as High-Altitude Pseudo-Satellites (HAPS), include
military surveillance, civilian research, observation, and acting as a radio communi-
cations relay, among others.
The first point to make when starting a new technology roadmap is that each
technology roadmap should have a clear and unique identifier and name:

8.2.1 2SEA – Solar-Electric Aircraft

This indicates that we are dealing with a “level 2” roadmap at the product level (see
Fig. 8.4), whereas “level 1” would indicate a market-level roadmap and “level 3” or
“level 4” would indicate an individual technology roadmap at the subsystem or
component level.
Next, the technology roadmap needs an outline or “table of contents.” Many
technology roadmaps only consist of a single slide or page (similar to Figs. 8.1, 8.2,
8.3 and 8.4). However, this is usually not sufficient to rationalize, quantify, and
explain the recommendations made by the roadmap. Here, we propose the follow-
ing outline for 2SEA3:

3
These 12 elements are a general recommendation for the outline and content of a technology
roadmap. In our technology roadmapping and development class at MIT, we follow this outline
224 8 Technology Roadmapping

1. Roadmap overview.
2. DSM allocation (interdependencies with others roadmaps).
3. Roadmap model (e.g., using OPM ISO 19450).
4. Figures of merit (FOM): Definition, name, unit, and trends dFOM/dt.
5. Alignment with company strategic drivers: FOM targets.
6. Positioning of company vs. competition: FOM charts.
7. Technical model: Morphological matrix and tradespace.
8. Financial model: Technology value (∆NPV).
9. Portfolio of R&D projects and prototypes.
10. Keys publications, presentations, and patents.
11. Technology strategy statement (incl. “arrow” or “swoosh” chart).
12. Roadmap maturity assessment (optional).
We now demonstrate what these elements might look like for the 2SEA roadmap.
1. Roadmap Overview
Solar-electric aircraft are built from lightweight materials such as carbon-fiber
reinforced polymers (CFRP) and harvest solar energy through the photoelectric
effect by bonding thin-film solar cells to the surface of the main wings, and poten-
tially the fuselage and empennage as well. The electrical energy harvested during
the day is then stored in onboard chemical batteries (e.g., lithium-ion, lithium-­
sulfur, etc.) or regenerative fuel cells and used for propelling the aircraft at all times,
including at night. For the system to work, there needs to be an overproduction of
energy during the day, so that the aircraft can use the stored energy to stay aloft at
night. The flight altitude of about 60,000–70,000 feet is critical to staying above the
clouds and not to interfere with commercial air traffic. Depending on the length of
day, that is, the diurnal cycle that determines the number of sunshine hours per day,
which itself depends on the latitude and time of year (seasonality), the problem is
easier or harder. The reference case in this technology roadmap is an equatorial mis-
sion (latitude = zero) with 12 hours of day and 12 hours of night.
The working principle and architecture of a typical solar-electric aircraft are
depicted in Fig. 8.7. Such diagrams are helpful in depicting the key elements of a
technology.
2. Design Structure Matrix (DSM) Allocation
In a dependency structure matrix (DSM), also known as a design structure
matrix, we identify other roadmaps at the same or at other levels that are coupled to
this roadmap. The coupling can be due to coinvestment relationships where an R&D
project or demonstrator (prototype) requires progress in another technology as well.
Coupling also exists when competing (mutually exclusive) technologies are being
pursued at the same time, leading to an eventual down-select of the winning
technology.
The 2-SEA roadmap tree that we can extract from the DSM (Fig. 8.8 right)
shows us that the solar-electric aircraft (2SEA) roadmap is part of a larger company-­
wide initiative on electrification of flight (1ELE) and that it requires the following

and add between 15 and 20 technology roadmaps per year, see [Link]
8.2 Example of Technology Roadmap: Solar-Electric Aircraft 225

Fig. 8.7 Working principle and architecture of a typical solar-electric aircraft

Fig. 8.8 DSM links of the 2SEA roadmap to other roadmaps at other levels

key enabling technologies at the subsystem level: 3CFP carbon fiber polymers,
3HEP hybrid electric propulsion, and 3EPS nonpropulsive energy management
(e.g., this includes the management of the charge-discharge cycles of the batteries
during the day-night cycle).
In turn, these level 3 technologies require enabling technologies at level 4, the
technology component level: 4CMP components made from CFRP4 (spars, wing
box, and fairings), 4EMT electric machines (motors and generators), 4ENS energy
sources (such as thin-film photovoltaics bonded to flight surfaces), and 4STO
(energy storage in the form of lithium-type batteries or regenerative fuel cells). This
hierarchy of roadmaps and the DSM allows to view a technology roadmap not in
isolation but in the context of the higher-level (i.e., the market viewpoint), which
sets performance, cost, safety, and reliability targets, and the lower-level more

4
CFRP = carbon fiber reinforced polymers.
226 8 Technology Roadmapping

Fig. 8.9 Object-process diagram (OPD) of the 2SEA solar-electric roadmap

detailed technology roadmaps that contain the enabling and supporting technolo-
gies needed to achieve the higher-level targets.
3. Roadmap Model Using Object-Process Methodology (OPM)
An important aspect of technology roadmapping is to clearly define the scope of
the technology covered by the roadmap. This sounds simple, but in practice may not
always be so clear. For example, does a roadmap on “high power electronics”
include only switches (e.g., MOSFETs) or does it also contain the filters, cables,
and control software? In this spirit, we provide an object-process diagram (OPD)5
of the 2SEA roadmap in Fig. 8.9.
This diagram captures the main object of the roadmap (solar-electric aircraft), its
various instances including the main competitors, its decomposition into subsys-
tems (wing, battery, e-motor, etc.), its characterization by figures of merit (FOMs),
as well as the main processes (flying and recharging).
An object-process language (OPL) description of the roadmap scope is auto-­
generated and given in the Appendix. It reflects the same content as Fig. 8.9, but in
a formal natural language. While initially awkward for the uninitiated, this kind of
semantically rigorous and formal description helps avoid unnecessary ambiguities
and confusion in terms of technology roadmap scope.
4. Figures of Merit (FOM) Definition
The roadmap should also be unambiguous when it comes to the figures of merit
(FOMs) that will be used to establish the status quo of the technology, its historical
trends, and where it should be heading in the future. Table 8.2 shows a list of FOMs
by which solar electric aircraft can be assessed. The first four (shown in bold) are
used to assess the aircraft itself. They are very similar to the FOMs that are used to
compare traditional aircraft that are propelled by fossil fuels. The big difference is
that 2SEA is emission free during flight operations.

5
OPD and OPL are based on ISO Standard 19,450 (2015) for object-process methodology (OPM).

Licensed to ([Link]@[Link])
8.2 Example of Technology Roadmap: Solar-Electric Aircraft 227

Table 8.2 FOMs for 2SEA solar-electric aircraft roadmap

FOM name Units Description
Unit Cost [€] Unit cost to manufacture the aircraft (incl. amortization of
R&D)
Operating Cost [€/FH] Cost per flight-hour, including all variable cost (e.g. energy
recharging, battery replacement), and maintenance
Maximum Payload [kg] Useful payload that can be carried (includes cargo, payload,
sensors and comm equipment) and passengers
Endurance [hrs] Time-aloft without recharging on the ground
Energy Storage [kWh/ Energy stored onboard per unit mass of energy storage devices
Density kg] (e.g. batteries)
Recharging Rate [kWh/ Rate at which batteries can be recharged on the ground
hr]
Electrical Max [kW] Total maximum electrical power generated on board by
Power e-machines, for both propulsive and non-propulsive use
Photovoltaic Cell [%] Conversion efficiency from incoming photon flux to useable
Efficiency electric current (electron flux)
Availability [hrs/y] Expected number of flight hours the aircraft is available for
service per year (excludes maintenance downtime)

The other rows in Table 8.2 represent subordinated FOMs which impact the per-
formance and cost of solar electric aircraft, but are provided as outputs (primary
FOMs) from lower-level roadmaps at level 3 or level 4, see Fig. 8.8.
Besides defining what the FOMs are, this section of the roadmap should also
contain the FOM trends over time dFOM/dt as well as some of the key governing
equations that underpin the technology. These governing equations can be derived
from physics (or chemistry, biology, etc..) or they can be empirically derived from a
multivariate regression model.6 Fig. 8.10 shows an example of a key governing
equation governing (solar-) electric aircraft.
The equation shown here is the electric version of the famous Bréguet range
equation (which will be introduced in Chap. 9) and estimates the all-electric range
as a function of key aerodynamic, structural, and electrical parameters. Some of the
improvement trends for photovoltaic cells were shown in Chap. 4. For example,
single crystalline silicon cells have been improving at a rate of about +0.4% per
year, but are subject to a maximum theoretical efficiency bound of 33.16%.
5. Alignment with Company Strategic Drivers
This section of the roadmap creates a link between the market-facing strategies
of the company (the top two layers shown in Fig. 8.5): Market and business and the
product-level FOMs and targets that should be achieved. Note that the analysis of
current and evolving markets and the setting of the business strategy is not part of
technology roadmapping, but feeds into it.

6
In general, physics-based models are preferred since empirically derived models are only valid
over the interval of training data that were used on the input side. As technology progresses, the
correlations derived for the empirical models may no longer be valid.
228 8 Technology Roadmapping

Fig. 8.10 Governing equation with inputs and outputs for (solar-) electric aircraft

Table 8.3 Strategic drivers for the 2SEA roadmap and statements of alignment
Number Strategic driver Alignment and targets
1 To develop a multipurpose solar-powered The 2SEA technology roadmap will target
HAPS (UAV) that has enough endurance a solar-powered UAV with a useful
and payload to provide a new payload of at least 10 kg and an endurance
commercially viable service that will of 500 days. This driver is currently
generate $X million in revenue by 2030 aligned with the 2SEA technology
roadmap
2 To develop autonomous flight capabilities The 2SEA technology roadmap will help
for HAPS and low Earth orbit (LEO) develop and test a certifiable stack of
satellites that will avoid the need for autonomy software that will reduce the
dedicated ground stations operating cost compared to current UAVs
by 50%. This driver is currently not
aligned with 2SEA

Table 8.3 shows an example of potential strategic drivers and alignment of the
2SEA technology roadmap with it.7
The list of drivers shows that the company views HAPS as a potential new busi-
ness and wants to develop it as a commercially viable (for profit) business (1). In
order to do so, the technology roadmap performs some analysis – using the govern-
ing equations in the previous section – and formulates a set of FOM targets that state
that such a UAV needs to achieve an endurance of 500 days (as opposed to the world
record of 26 days that was demonstrated in 2018) and should be able to carry a
payload of 10 kg. The roadmap confirms that it is aligned with this driver. This
means that the analysis, technology targets, and R&D projects contained in the
roadmap (and hopefully funded by the R&D budget) support the strategic ambition
stated by driver 1. The second driver, however, which is to use the HAPS program
as a platform for developing an autonomy stack for both UAVs and satellites, is not
currently aligned with the roadmap.8

7
Disclaimer: While we have used the Zephyr as a motivating example at the beginning of this sec-
tion, the strategic drivers in this section should not be taken as a direct reflection of the Airbus
Defense and Space business strategy in the area of solar electric aircraft.
8
Not all targets or ambitions stated in a technology roadmap may initially be funded or fundable
by the R&D budget. That is fundamentally okay, since the technology roadmap is a statement of
ambitions, translated to quantified targets. However, once converged, the technology roadmap tar-
gets should be achievable both fiscally and in terms of their feasibility within physical limits.
8.2 Example of Technology Roadmap: Solar-Electric Aircraft 229

As can be seen in Fig. 8.8, there are currently no autonomy or software-related

elements in this roadmap. Therefore, the roadmap has an internal conflict between
the strategic ambition expressed by strategic driver 2 and what is actually planned
in the roadmap. This conflict needs to be first acknowledged and ultimately resolved,
either by expanding the scope of the 2SEA roadmap or by explicitly removing stra-
tegic driver 2 as a requirement.
6. Positioning: Company Versus Competition FOM Charts
The next portion of the roadmap is a careful qualitative and quantitative bench-
marking of the company’s position against the present and potential future (if
known) competition in this particular segment. This benchmarking is best done via
a set of FOM charts so that the “gap” between the company and its competitors can
be quantified and visualized. In some FOMs, the company may be the leader, while
in others a follower. This task requires the gathering of data through the technology
scouting function (see Chap. 14 for details).
Figure 8.11 shows a summary of current and past electric and solar-electric air-
craft from public data.
This is an important exercise to bring some realism to the technology roadmap in
terms of already fielded products, services, and systems and those under develop-
ment. The aerobatic aircraft Extra 330LE by Siemens had the world record for the
most powerful flight-certified electric motor (260 kW) at the time of writing. The
Pipistrel Alpha Electro is a small electric training aircraft which is not solar pow-
ered, but it is in serial production. The Zephyr 7 is the previous version of Zephyr
which established the prior endurance world record for solar-electric aircraft
(14 days) in 2010. The Solar Impulse 2 was a single-piloted solar-powered aircraft
that circumnavigated the globe in 2015–2016 in 17 stages, the longest being the one
from Japan to Hawaii (118 hours).
SolarEagle9 and Solara 50 were both very ambitious projects that aimed to launch
solar-electric aircraft with very aggressive targets (endurance up to 5 years) and

Fig. 8.11 Benchmarking of (solar-) electric aircraft (approximations are made where necessary)

9
This project was partially funded by the DARPA Vulture program whose aim it was to develop a
solar-powered UAV that could fly for 5 years without landing. The project was canceled in 2012.
230 8 Technology Roadmapping

Fig. 8.12 Endurance [hrs] versus payload [kg] for all-electric and solar-electric aircraft

payloads up to 450 kg. Both of these projects were canceled prematurely. Why
is that?
The answer is shown in Fig. 8.12.
The Pareto front (see Chap. 4, Fig. 4.17 for a definition) shown in black in the
lower-left corner of the graph shows the best trade-off between endurance and pay-
load for actually achieved electric flights by 2017. The Airbus Zephyr, Solar Impulse
2, and Pipistrel Alpha Electro all have certified flight records that anchor their posi-
tion on this FOM chart. It is interesting to note that Solar Impulse 2 overheated its
battery pack during its longest leg in 2015–2016 and, therefore, pushed the limits of
battery technology available at that time. We can now see that both Solar Eagle in
the upper right corner and Solara 50 in the upper left corner were chasing FOM
targets that were unachievable with the technology available at that time.
The progression of the Pareto front shown in red corresponds to what might be a
realistic Pareto front progression between 2017 and 2020. Airbus Zephyr Next-­
Generation (NG) has already shown with its world record (624 hours endurance)
that the upper left target (low payload mass of about 5 kg and high endurance of
600+ hours) is feasible. There are currently no plans for a Solar Impulse 3, which
would be a non-stop solar-electric circumnavigation of Earth with one pilot, and
which would require a nonstop flight of about 450 hours. A next-generation E-Fan
aircraft with an endurance of about 2.5 hours (all electric) also seems within reach
for 2020. Then, in green we set a potentially more ambitious target Pareto front for
2030. This is the ambition of the 2SEA technology roadmap as expressed by strate-
gic driver 1.
We see that in the upper left the Solara 50 project, which was started by Titan
Aerospace, and then acquired by Google, then cancelled, and which ran from about
2013 to 2017, had the right target for about a 2030 entry into service (EIS), but not
8.2 Example of Technology Roadmap: Solar-Electric Aircraft 231

for 2020 or sooner. The target set by Solar Eagle was even more utopian and may
not be achievable before 2050 according to this 2SEA roadmap.
The positioning (where are we today?), benchmarking (where is our competi-
tion?), and target setting (where do we want to be in 2 years? 5 years? 10 years?)
and Pareto front progression are an essential part of a technology roadmap.
It is this kind of information that allows technical leaders to push back against
unrealistic business targets and to set the right expectations. The existence of this
kind of quantitative and validated information is what distinguishes useful and high-­
quality roadmaps from “pseudo-roadmaps” that are mainly qualitative in nature and
primarily useful as a visual aid (usually in the form of a PowerPoint chart) or con-
ceptual guideline but not for detailed and serious technical planning. More on this
topic in Sect. 8.5 on Technology Roadmapping Maturity Levels below.
7. Technical Model
In order to assess the feasibility of technical (and financial) targets at the level of
the 2SEA roadmap, it is necessary to develop a technical model. The purpose of
such a model is to explore the design tradespace and establish what are the active
constraints in the system. The first step can be to establish a morphological matrix
that shows the main technology selection alternatives that exist at the first level of
decomposition, see Fig. 8.13.
It is interesting to note that the architecture and technology selections for the
three aircraft on the 2017 Pareto front (Zephyr, Solar Impulse 2, and E-Fan 2.0) are
quite different. While Zephyr uses lithium-sulfur batteries, the other two use the
more conventional lithium-ion batteries. Solar Impulse uses the less efficient (but
more affordable) single-cell silicon-based photovoltaics, while Zephyr uses spe-
cially manufactured thin-film multijunction cells.
The technical model centers on the E-range and E-endurance equations and com-
pares different aircraft sizing (e.g., wingspan, engine power, and battery capacity)
taking into account aerodynamics, weights and balance, the performance of the air-
craft, and also its manufacturing cost. It is recommended to use multidisciplinary
design optimization (MDO) when selecting and sizing technologies in order to get
the most out of them and to compare them fairly (Fig. 8.14).
8. Financial Model
While technology roadmapping can also be important for not-for-profit enter-
prises, such as the NASA technology roadmaps discussed in Sect. 8.3, it is essential
in a technology-based for-profit business. How much should the company expect to
spend on R&D and on what projects? What % improvement in key FOMs can be
expected and by when? How much are customers willing to pay for such improve-
ments? How much internal cost reduction can be achieved due to new technologies
(see also Chap. 12)?
A financial model is akin to a “business plan,” but not necessarily for the product as
a whole, but for the “delta” or relative impact that a specific technology can have on a
baseline business plan. Imagine that a business plan for a product includes only well-
established technologies. How would the business plan change with the “new”
232 8 Technology Roadmapping

Fig. 8.13 Morphological matrix for (solar-) electric aircraft

Fig. 8.14 Multidisciplinary design optimization model of solar-electric aircraft

technology included? Would it be better or worse? How would the uncertainty of the
business plan (standard deviation of net present value (NPV)) be affected by the
technology?
Figure 8.15 contains a sample NPV analysis underlying the 2SEA roadmap. It
shows the nonrecurring cost (PDP NRC) of the product development project, which
8.2 Example of Technology Roadmap: Solar-Electric Aircraft 233

Fig. 8.15 Hypothetical financial model for the 2SEA roadmap, PDP NRC Product Develeopment
Project Non Recurring Cost, MFG RC Manufacturing Recurring Cost

includes the R&D expenditures as negative numbers. A ramp-up period of 4 years

is planned with a flat revenue plateau (of 400 M€ per year) and a total program dura-
tion of 24 years.
Such a model can then be used as a mechanism for quantifying the relative
impact of different technologies such as extended endurance and larger payload
(more revenue) or using a more expensive battery technology (more cost per unit).
Ultimately, the decisions on product and service launch (or not) and which tech-
nologies to incorporate are linked to specific financial and technical FOM-based
targets and milestones on a timeline and are the decisions of the senior management
(and in some cases the board) of the company.
9. Portfolio of R&D Projects and Prototypes
In order to know whether the technology roadmap will be able to meet its targets,
it is necessary to define and prioritize which R&D projects should be carried out and
234 8 Technology Roadmapping

funded in the overall portfolio. This is an important section of the technology road-
map since it creates a link between the higher-level financial and technical FOM-­
based targets and the specific R&D activities and projects that the technical
organizations (research centers, R&D departments, engineering, etc.) will carry out,
either internally or in collaboration with partners.
In order to select and prioritize R&D projects, we recommend using the techni-
cal and financial models developed as part of the roadmap to rank-order projects
based on an objective set of criteria and analyses.10 Figure 8.16 illustrates how tech-
nical models can be used to make technology project selections, for example, based
on the previously stated 2030 performance targets (see Fig. 8.12). Figure 8.17 shows
the outcome if none of the three potential R&D projects is selected.
This model makes an important assumption: even if the company decides not to
invest in any of the three proposed projects (battery, solar cell, and structural
improvements), those technologies will still progress “on their own.” This is due to
the fact – as shown in Chap. 4 – that long-term technology improvement trends are
quite predictable and that, for most or at least for many technologies, there are sev-
eral competing players and suppliers around the world.
Major technological improvements are almost never achieved by just one com-
pany or organization (despite some claims made by these firms or the media) and
often rely on a complex web of contributions from many organizations and
individuals.

Fig. 8.16 R&D project evaluation using a product-specific technical model

10
In many organizations, R&D projects are selected based mainly on “intuition” alone and the
voices of a few – usually senior and very experienced – individuals. This is potentially a dangerous
way to go as Christensen shows (Chap. 7) due to the innovator’s dilemma. Usually this intuition-
based process by entrenched senior engineers and executives will favor sustaining incremental
technology investments, instead of sustaining radical or even disruptive ones. The dynamics and
pitfalls of R&D project selection and R&D portfolio management are discussed further in Chap. 16.
8.2 Example of Technology Roadmap: Solar-Electric Aircraft 235

Fig. 8.17 Expected outcome if none of the three proposed R&D projects are selected

So, for the owner of the 2SEA roadmap, the fundamental question is: “Can I sit
back and wait until my subsystem and component technologies have matured ‘natu-
rally’ based on their expected ‘natural’ rate of progression (the solid blue lines in
Fig. 8.16 left), or do I need to proactively invest in them to remain or become a
leader and accelerate their development (the dashed red lines in Fig. 8.16 left)?”
For the 2030 target set in Fig. 8.17 (right), the answer is clear: We are unable to
meet the target with no R&D investments in individual technologies. If we scale
back the target to a payload of less than 10 kg and an endurance of less than 500 days,
the target could potentially be met. We now consider investing in each project, one
at a time, as shown in Fig. 8.18.
The results of the analysis show that the largest impact on the performance of the
aircraft is the battery technology (in this case, using lithium-sulfur chemistry). This
makes sense since at its current size the aircraft is able to generate enough electrical
power during the day (at least at an equatorial latitude and 12-hour day); however,
it is its ability to store and release this energy efficiently at night in terms of energy
density [J/kg] where a large improvement is needed. The problem is compounded
by the deterioration of the battery with each cycle. Interestingly, further improving
solar cell efficiency has no impact since it is not an active constraint in the system.
Also, structural improvements alone (lightweighting of the structure) are insuffi-
cient. A further analysis would look at the net effect of combinations of different
projects and technologies (this will be further discussed in Chap. 16 on R&D port-
folio management).
For now, the company decides on two projects in the 2SEA roadmap:
1. A Li-S battery improvement project with the FOM target of raising the number
of charge-discharge cycles from 100 to 500 by 2025. This project will be allo-
cated to the linked 4STO Energy Storage Roadmap and executed with a partner
who specializes in lithium-sulfur chemistry-based battery development and cer-
tification (with shared IP, see Chap. 5).
236 8 Technology Roadmapping

Fig. 8.18 Impact of individual R&D project investments on the Pareto frontier: top: Li-S battery
improvements alone, middle: solar cell efficiency improvements alone, and bottom: structural
improvements alone

2. A flight demonstrator project will be launched as part of the 2SEA roadmap to

demonstrate a 10 kg payload and 365-day (1 full year) capability by 2027 as a
prototype, with an intended entry into service (EIS) of a commercial 500-day-­10-
kg-capable product and associated profitable service by 2030.
10. Key Publications, Presentations, and Patents.
A technology roadmap should contain a comprehensive list of publications, pre-
sentations, and key patents as shown in Fig. 8.19. This includes literature trends,
papers published at key conferences and in the trade literature, and the trade

Licensed to ([Link]@[Link])
8.2 Example of Technology Roadmap: Solar-Electric Aircraft 237

Fig. 8.19 Key scientific publications, trade press summaries, patent analysis, and publication
trends should be included in a high-quality technology roadmap

press. Depending on legal considerations the technology roadmap may or may not
contain patent information (since this could affect potential discovery in a future
infringement lawsuit).
Given the continual nature of technology progress, a technology roadmap cannot
be created once and then left unattended for long periods of time. The best-in class
companies that use technology roadmapping effectively have dedicated roadmap
owners. This can be a full time or part-time job depending on the complexity and
strategic importance of the roadmap. Technology roadmaps need to be refreshed
regularly. Refresh rates of technology roadmaps depend on the industry and the
dynamics of innovation. A yearly refresh that is synchronized with the annual plan-
ning and budget cycle is the minimum refresh rate that should be expected for tech-
nology roadmaps.
11. Technology Strategy Statement
A technology roadmap should conclude and be summarized by both a written
statement that summarizes the technology strategy coming out of the roadmap as
well as a graphic that shows the key R&D investments, targets, and a vision for this
technology (and associated product or service) over time. The technology roadmap
could also insert a “swoosh” chart at this point. A maturity assessment of the road-
map (section 12) is optional, but recommended. For the 2SEA roadmap, the state-
ment could read as follows:
Our target is to develop a new solar-powered and electrically driven UAV as a
HAPS service platform with an entry-into-service date of 2030. To achieve the target
of an endurance of 500 days and useful payload of 10 kg, we will invest in two R&D
projects. The first is a flight demonstrator with a first flight by 2027 to demonstrate
a full-year aloft (365 days) at an equatorial latitude with a payload of 10 kg. The
second project is an accelerated development of Li-S batteries with our partner XYZ
with a target lifetime performance of 500 charge-discharge cycles by 2027. This is
an enabling technology to reach our 2030 technical and business targets.
238 8 Technology Roadmapping

8.3 NASA’s Technology Roadmaps (TA1–15)

Technology roadmaps are not only in use in the industrial (for profit) sector.
One of the organizations that has developed and made extensive use of technol-
ogy roadmaps is the National Aeronautics and Space Administration (NASA) in the
United States. There was a major effort in the agency to create an initial set of road-
maps in 2012. These were then updated in 2015 and decomposed into the 15 techni-
cal areas (TAs) shown in Fig. 8.20.
One interesting fact about the NASA technology roadmaps is that when they
were first published in 2012 that only TA1-TA14 existed. In other words, the tech-
nology roadmaps focused only on technologies related to human and robotic space
missions. Later, in 2015, the TA15 roadmap was added which includes all of
aeronautics.
Given the breadth of NASA’s missions and activities, each of these roadmaps
contains many levels of decomposition in order to capture comprehensively the
technological base. Let’s consider as an example area TA9 which covers entry,
descent and landing (EDL) systems. Inside the roadmap we find three levels of
technology decomposition as shown in Fig. 8.21.
Within the roadmap we can then deep dive into a set of missions that create the
“technology pull” or “need” for new or enhanced technologies. For example,
Fig. 8.22 shows a “Venus In-Situ Explorer” as a potential mission with an originally

Fig. 8.20 NASA’s technology roadmaps grouped into 15 technical areas (TAs)
8.3 NASA’s Technology Roadmaps (TA1–15) 239

Fig. 8.21 Decomposition inside the NASA TA9 EDL Roadmap

Fig. 8.22 Timeline for TA9 technology roadmap from missions to technologies
240 8 Technology Roadmapping

Fig. 8.23 Illustration of Rigid Venus Entry Probe (left) and a Mechanically-Deployable
Aeroshell (right)

planned launch date in 2024 (green triangle on the upper right). Entering the atmo-
sphere of Venus which is hotter and denser than the atmosphere of Earth or Mars
will require a heat shield that can withstand the thermal loading during ballistic re-­
entry as shown in Fig. 8.23.
The timeline then backtracks from the planned mission launch date (2024) to the
point of “need,” where key technologies should be available at a given TRL level.
For example, the technologies under 9.1.1. and 9.1.2. (thermal protection systems
for rigid and deployable decelerators) are shown to be “needed” by 2016 and should
start development in 2014. The rule of thumb is that technologies should have been
matured to at least TRL 6 before they are taken onboard by a flight program. A simi-
lar rule of thumb exists in the commercial sector as well.
Figure 8.24 provides a description of the technology and its key challenges (top),
quantification of the current technological state of the art (left), technology perfor-
mance goal (right), and technology interdependencies on research or on other tech-
nologies (bottom). We see here that the FOM-based targets for this technology are
ambitious but not utopian: a peak heating rate of 50–100 [W/cm2] (a factor of 2
improvement), an integrated heat load during reentry of [12 kJ/cm2] (a factor > 2
improvement), a peak temperature of 400 degrees [C] (a 30% improvement), and a
deployed diameter of 10–25 meters (a factor of 2–4 improvement). Clearly, this
technology must be considered as an enabling technology if we desire to enter the
atmosphere of Venus at high speeds.
How are these Technology Roadmaps Used at NASA?
Figure 8.25 depicts the nominal NASA technology roadmapping process. The
TA1–15 roadmaps are shown on the upper left. They serve as input to the NASA
Technology Executive Council (NTEC). This is the decision body that sets technol-
ogy policy, and prioritizes strategic technology investments. The roadmaps contain
a larger “wishlist” than what can be funded (this is usually true in all organizations)
and so a down-selection and prioritization is necessary. This then influences NASA’s
annual budget process, leading to a certain number of technology projects that are
8.3 NASA’s Technology Roadmaps (TA1–15) 241

Fig. 8.24 Technology description and performance goal for EDL heat shield technology

Fig. 8.25 NASA

technology
roadmapping and budget
process

funded. These projects are then documented and the portfolio is analyzed and
reflected in “TechPort” which is an agency-wide technology database.
A portion of this database known as “Tech Finder” is then made available to the
public, containing information on patents, licenses, and software agreements. The
loop is then closed periodically by injecting new information into the roadmaps
from TechPort.
242 8 Technology Roadmapping

⇨ Exercise 8.1
Pick a published roadmap, for example, from NASA or any other you can
obtain (however, do not use or share company confidential materials). Perform
a careful review of the roadmap and critique it on 1–2 pages. What technology
is it about? What elements does it contain? Is anything missing according to
the proposed outline? Is it fit for purpose?

Since 2016, NASA has gone to a more decentralized approach to technology

planning with every directorate and major program setting their own priorities for
technology investment, usually based on the strategic input coming from decadal
surveys compiled by the National Research Council (NRC) of the National
Academies of Science, Engineering and Medicine.11

8.4 Advanced Technology Roadmap Architecture (ATRA)

This section briefly describes the technology planning and roadmapping approach
as it was implemented and refined at a major aerospace company by the author and
his team. It is based on the principles of technology roadmapping described in this
book. Figure 8.26 shows the overall ATRA methodology and its four major steps.
The inputs to the ATRA methodology are as follows and are shown on the left
side of Fig. 8.26:
• A hierarchical decomposition of the product, service, and technology portfolio
into different mapped levels. The simplest decomposition is one with two levels
with products (and services or missions) at level 1 and technologies at level 2
(see also Fig. 8.1). A more fine-grained decomposition was shown in Fig. 8.8
with four levels of decomposition: markets or missions (L1), products and ser-
vices (L2), subsystems (L3), and components (L4).
• Based on the DSM of each individual roadmap (see Fig. 8.8), which shows the
interdependencies with other products and technologies (a technology can and
should ideally serve more than one product), a global Dependency Structure
Matrix (DSM) can be constructed which shows an overview of the total system
of roadmaps, potentially including the selected R&D projects.
• Strategic drivers coming from marketing, strategy, and senior management. See
Table 8.3 for an example of strategic drivers.
• Other inputs such as those coming from technology scouting, IP analytics, and
subject matter experts (SMEs) both inside and outside the company.

11
NASA has recently selected the ATRA framework for researching improved ways of managing
its technology portfolio, see: [Link]
tions-esi/esi2020/astra/
8.4 Advanced Technology Roadmap Architecture (ATRA) 243

Fig. 8.26 Advanced technology roadmap architecture (ATRA)

The ATRA methodology then proceeds in four steps (see Fig. 8.26 middle col-
umn), each asking a very specific question that must be answered by the roadmap.
1. Where Are We Today?
This question asks for the current status quo in terms of market position, prod-
ucts, services, technology performance (FOM-based), and running R&D projects.
The corresponding sections in the technology roadmaps that capture the status quo
are several as demonstrated in the 2SEA example.
When starting technology roadmapping from scratch or building on a rather thin
initial set of roadmaps, this can be a rather laborious process, involving several
workshops (Fig. 8.27), and potentially dozens or even hundreds of stakeholders in
the organization.
Depending on the size of the organization and the complexity of its product and
service portfolio, the set of R&D projects, and the number of subject matter experts
involved, this can yield thousands of pieces of information that need to be collated,
grouped, linked, and validated. In some cases, there may be ambiguities in terms of
which roadmap a project belongs to or what is the primary product in need of a
particular technology.
Given this initial set of information, roadmap owners (RMOs) or technology
committees are then appointed to develop the individual roadmaps. The content of
the roadmaps should be in a more or less standardized format. One of the key out-
puts of step 1 is a set of FOM charts, as shown in Fig. 8.26 (top right). It shows the
current position(s) of the company compared to its competitors and compared to the
current state of the art (SOA), expressed as a Pareto frontier. See Fig. 8.12 for a
quantitative example in the 2SEA roadmap.
This will give a clear sense in which technology areas the company is leading,
where it is about equal to its peers, and where it is behind its competitors.
244 8 Technology Roadmapping

Fig. 8.27 Interactive and hands-on workshops are recommended to map all the running R&D
projects against the technology roadmaps, subject matter experts, including roadmap owners, and
against target products and services (or missions)

2. Where Could We Go?

The second step asks the question of what are the possible new products, ser-
vices, or technologies that the company could pursue. Some of these could be based
on a market or customer “pull,” while some could be based on technology “push”
based on the ideas of the science and engineering community or the leadership
within the company. The existence of a so-called Concurrent Design Facility (CDF)
is very helpful as a central focal point and supporting infrastructure for this explor-
atory activity (Fig. 8.28).
This is an important phase of roadmapping which is often neglected or cut short.
It is important to explore new concepts, missions, and ideas, and to do so using both
qualitative concepts and quantitative models. An effective methodology during this
stage is “Concurrent Engineering” (Knoll et al. 2018), where different experts from
inside and outside the organization are brought together to explore potential techno-
logical directions.
As shown in Fig. 8.26 (middle), we would expect a discrete set of potential scenar-
ios for the evolution of existing products and technologies to emerge from step 2 and
the associated exploratory CDF sessions. Figure 8.26 shows a scenario A and scenario
B which are initially common, but then split after about 3 years (decision point).
3. Where Should We Go?
This step applies a prioritization scheme to the proposed product and technology
scenarios. It is here that the specific FOM targets are set and agreed for individual
products and technologies. If these are very different from the original input received
from marketing or strategy, there may have to be an iteration loop to close any gaps
or inconsistencies.
8.4 Advanced Technology Roadmap Architecture (ATRA) 245

Fig. 8.28 Concurrent Design Facility (CDF) in support of technology roadmapping

Fig. 8.29 Network chart (left) coming from step 1 “as is” analysis, reorganized as a “DSM” in
steps 2 and 3 with clearly formed clusters of technologies (referred to as “thrusts” shown right)
which represent strategic investment areas for technologies

Figure 8.29 shows an example of the clarification that should come from steps 1
and 2 moving into step 3. We see on the left a network diagram that shows the inter-
dependencies between different roadmaps across the ATRA, with some technolo-
gies having a central role as enabling technologies for several products. On the
right, we see the same information, but now organized as a DSM with L1 products
in the upper left and L2 technologies on the lower right. The technologies are
grouped into technology clusters: digital design and manufacturing (DDM), materi-
als, autonomy, connectivity, and electrification as an example. These clusters of
roadmaps then become the focal points for targeted technological investments
including R&D projects and new prototypes and demonstrators.
246 8 Technology Roadmapping

4. Where We Are Going!

This is the final step in technology roadmapping and technology planning. This
is where the set of proposed scenarios and R&D projects and new demonstrators
has to be fitted into an overall budget envelope for R&D. Most companies spend
anywhere between 1% and 20% of their revenues on R&D to “prepare for the
future.”
The goal of technology roadmapping is to turn this difficult process from a more
intuitive and personality-driven exercise into a disciplined and rational one. At this
stage, each potential project in the R&D portfolio should have clear targets, a state-
ment of work (SOW), and a well fleshed out budget to completion.
To build a competitive R&D portfolio, following the recommendations of the
roadmaps, and hopefully achieving a consensus among the senior technical leader-
ship, the following decisions have to be made on a project-by-project basis:
• START: Which new technology projects should be started and with what targets?
• STOP: Which projects are coming to completion naturally (handoff to product
development) and which projects should be terminated prematurely either
because they have stalled or because their deliverables are no longer needed due
to a change in strategic drivers?
• KEEP: Continue funding projects that are on track and that are still needed.
• CHANGE: Modify ongoing projects for a variety of reasons. For example, proj-
ects may be accelerated or slowed down depending on changes in product
entry-into-service (EIS) dates. Or similar projects in the portfolio may be merged
to achieve better synergies.
The ability to stop or change projects may be limited by external commitments
that have been made such as collaborations with development partners on a given
project or external funding agencies. Figure 8.30 shows an example of the type of
technology projects that have been funded and rationalized in the R&D portfolio of
a major aerospace company based on the ATRA methodology.

Fig. 8.30 Sample of projects recommended by technology roadmaps in the R&D portfolio of a
major aerospace company. These projects line up with the technology clusters in Fig. 8.29 (right)

Licensed to ([Link]@[Link])
8.5 Maturity Scale for Technology Roadmapping 247

8.5 Maturity Scale for Technology Roadmapping

As stated by Phaal and Muller (2009), technology roadmapping is not really new. It
has been practiced since about the 1970s. However, the speed of technology devel-
opment, the number of companies that have been disrupted (see Chap. 7) or that
have disappeared due to a lack of technological investment or foresight, has
increased sharply in recent decades.
As a result, technology planning and roadmapping are now viewed as a key stra-
tegic function in many technology-intensive industries such as aerospace, automo-
tive, consumer electronics, software, life sciences, medical devices, and many more.
Recently, Schimpf and Abele (2019) have conducted an empirical survey of tech-
nology roadmapping in N = 81 German industrial firms, including smaller and mid-­
sized ones. Figure 8.31 shows a quantitative result from their survey in terms of the
mentioned application areas of roadmapping.
They found that:

*Quote
“Companies apply roadmapping within an average of 3.37 application areas with a
standard deviation of 1.17 and roughly one-third (32.1%) of participants apply road-
mapping for two application areas or less. This leads to a rejection of hypothesis
H01, recognizing that a majority of participating companies apply roadmapping to
more than two application areas. Within the content of roadmaps in companies,
products (79.7%), technologies (68.4%) and projects (57.0%) are the most common
options mentioned by participants.”

This confirms also the soundness of the ATRA approach which emphasizes a
clear mapping from products to technologies to projects. While roadmapping is
becoming more common among technologically intensive companies, the quality
and impact of technology roadmaps can vary greatly.

Fig. 8.31 Frequency of application areas for roadmapping in German companies (N = 81; multi-
ple responses possible). Source: Schimpf and Abele (2019)
248 8 Technology Roadmapping

Table 8.4 Technology Roadmapping Maturity Scale

Maturity
level Name Characteristics at that level
I Exploration Partial list of only the most important technologies
Focus mainly on technology scouting and finding “blind spots”
Uneven format, quality, and depth of roadmaps
Not used at all for decision-making, just for information
II Canvassing Complete list of roadmaps across the firm
Centralized project inventory mapped to roadmaps
Standardization of format and dedicated roadmap owners
“Flat” list of technologies, no explicit link to products
III Evaluation Explicit hierarchy of roadmaps with link to products or missions
Clear definition of FOMs and setting of targets
Anticipated entry-in-service dates are used to set pace
Find and exploit synergies across business units
IV Prescription Roadmaps are the main way to decide on R&D investments
Value for money is calculated for sustaining technologies
Quantified route-to-target options (vector charts) evaluated with risk
levels in products where multiple technologies are used
Clearly prioritized and ranked list of R&D projects in each roadmap
V Optimization Calculation of FOM targets and value with calibrated technical
models for each product, including the mapped technologies
Validated multiyear cost models for NRC and RC
Prioritization of R&D investments across product divisions
Portfolio optimization for value versus risk to maximize the NPV for
the firm with explicit expectations on ROI of the R&D portfolio

In Table 8.4, we propose a five-level maturity scale to assess technology road-

mapping in a given organization. The higher the level, the more advanced technol-
ogy roadmapping is practiced.
A company that is new to technology roadmapping should expect to start at level
I and – with the proper support (incl. financial resources) of the senior management
should be able to – progress about one level per year.
Thus, the full roadmapping journey from level I to level V may realistically take
5 years or more. It is also possible for a company to achieve a high level of maturity
in technology roadmapping at one point, but to regress again for a variety of rea-
sons, such as changes in management, lack of support from senior management, or
due to mergers and acquisitions of companies at vastly different levels of technol-
ogy roadmapping maturity one from the other.

⇨ Exercise 8.2
Develop a technology roadmap for a technology of your choice. Make sure
you are passionate about the technology you choose. This can be a quick
exercise to arrive at a sketch of a roadmap, or a big effort over multiple weeks
 Appendix 249

or months. Use 2SEA as an example for the format of the roadmap, but feel
free to add, modify, or remove elements as you see fit. Summarize your road-
map in a document or digital wiki (including the use of hyperlinks between
the elements) and present it to your peers or management for feedback.

Appendix

Object-process language (OPL) for the 2SEA roadmap

250 8 Technology Roadmapping

References

Bernal, Luis, et al. "Technology roadmapping handbook." International SEPT Program, University
of Leipzig (2009)
Kerr C, Phaal R. Technology roadmapping: Industrial roots, forgotten history and unknown ori-
gins. Technological Forecasting and Social Change. 2020 Jun 1;155:119967.
Knoll, Dominik, Alessandro Golkar, and Olivier de Weck. "A concurrent design approach
for model-based technology roadmapping." In 2018 Annual IEEE International Systems
Conference (SysCon), pp. 1-6. IEEE, 2018.
NASA Technology Roadmaps, Office of the Chief Technologist (OCT): [Link]
offices/oct/home/roadmaps/[Link]
Phaal, Robert, and Muller, Gerrit, “An architectural framework for roadmapping: Towards visual
strategy,” Technological Forecasting and Social Change, Volume 76, Issue 1, 2009,Pages
39-49, ISSN 0040-1625
Schimpf, Sven, and Thomas Abele. "How German Companies apply Roadmapping: Evidence from
an Empirical Study." Journal of Engineering and Technology Management, 52 (2019): 74-88.
Chapter 9
Case 2: The Aircraft

Advanced Technology Roadmap Architecture (ATRA)

Inputs
Steps Outputs
Strategic Drivers for Technology
+10y Technology

FOMjj
1. Where are we today? Roadmaps
L1 Products and Missions +5y

L2 Technologies Technology State of the Art and Organization Figures of Merit (FOM)
Competitive Benchmarking Current State of the Art (SOA)
Competitor 1 Technology Trends dFOM/dt
Technology Systems Modeling Competitor 2

Tech Pul
Pull Today FOMi
2. Where could we go?
Dependency Structure Matrix

L1
+10y
Scenario-based
Technology Systems Modeling and Scenario A Technology Valuation
FOMj

Trends over Time

Technology +5y Design Reference Missions
Scenario B
Projects Future Scenarios
? T h l
Technology V
Valuation
l ti
3. Where should we go? Vector Charts
L2
Scenario Analysis and FOMi
Technology Valuation
Tech Push Technology Investment
Efficient Frontier
E[NPV] - Return

Technology Scouting 4. Where we are going! Pareto-optimal set of technology

Knowledge Management Technology investment portfolios
Technology Portfolio Valuation, Portfolio Recommended Technology Portfolio
Intellectual Property Analytics Optimization and Selection Technology (Expected NPV and Risk)
Projects
σ[NPV] - Risk
Foundations C
Cases
9
Definitions History Nature Ecosystems The Future Case 1 Case 2 Case 3 Case 4
What is Milestones of Technology, Nature Technology Diffusion, Is there a Deep Space DNA
Singularity ? Automobiles Aircraft
Technology? Technology and Humans Infusion and Industry Network Sequencing

© Springer Nature Switzerland AG 2022 251

O. L. de Weck, Technology Roadmapping and Development,
[Link]