The evolution of technologies, industries and regions

In the earlier research on technological evolution in the 1970-1995 period, attention was mainly paid to either a whole economy or a single sector or technological paradigm. It is broadly understood from this research that different industries and technologies evolve at different rates. This means that over time, some industries may be more important than others, or at least, some may be accelerating while others may be stagnant or declining. In recent research by Saviotti and Pyka (2013), the emergence of new technologies and industries (and the goods and services that they provide) is seen as offsetting the diminishing returns that are innate in the development of existing technologies. Nelson (2015) argues that this is a reason why absorption and further development of these technologies are necessary to maintain economic development.

In enabling technological evolution in countries, a whole range of actors play a part. Individuals and informal networks, to large and small firms all play a role. However, for the last century, most technological advancements have been supported by scientists, the academia and professional societies and a range of supporting meso organisations. In Europe, professional associations often play an important role in the deepening and dissemination of technological knowledge.

I want to come back to the meso organisations mentioned in the earlier paragraph. Meso organisations or functions are created in response to structural issues like market failures, sometimes government failures or persistent patterns of underperformance in the economy. These meso functions are critical in supporting economic actors to discover what is possible in a given economic context, to assist stakeholders to overcome coordination failures, and to provide critical public goods (such as scarce or expensive technological infrastructure, demonstration facilities, testing facilities, public research, and so on).

The meso functions enable a society, industry or even the public sector to discover and absorb new ideas, they enable learning by doing, they encourage the adaptation and dissemination of new knowledge or technologies, and they connect different stakeholders to overcome coordination and search failures. These meso functions are a critical ingredient in the local innovation system as they extend the technological capability of a given sector, industry, market or region in a country.

You would have noticed that I have not yet mentioned universities and public research efforts. This is simply because I have written about them so often as they form a critical part of the local innovation system. I sometimes even think that the higher education sector receives too much attention. Yet, education from basic schooling to higher education plays a critical role. For me, a university is an important meso organisation, and research centres, technology extension centres and laboratories that provides testing facilities are all important meso functions or maybe even meso organisations hosted by a larger organisation.

The importance of the higher education sector in the technological infrastructure varies for different parts of the economy. Nelson contends that scientific and technological research and teaching, especially the more applied fields, provide a base of knowledge that is accessible to all technically sophisticated individuals and firms working to advance technology in a field (Nelson, 2015). However, different fields also depend, to different extents, on scientific and formal research and technology support. Therefore, measuring journal articles and research outputs as a contribution to the national innovation system or as a proxy for technological capability will always paint only a partial picture. It really also depends on the pace of change and scientific advancement that is taking place in a region, a technological domain or an industry.

Furthermore, different industries depend, to different extents, on government support and incentives. In some fields public support is crucial, and in other cases, provides little incentive or value. In many cases innovations preceded science, and continued development is only possible due to the iteration between researchers and enterprises. Nelson continues that the kinds of firms that do most of the innovating differ – in some fields this tends to be large, established firms while in others it is smaller firms or new start-ups (Nelson, 2015).

Nelson draws an important conclusion that has really shaped my own thinking. Nelson states that there is no single set of policies that are applicable to all technologies and industries. What will be effective in some fields will not be in others. For instance, small business promotion in some sectors in one country could work, but it could be ineffective in another country.

In South Africa, with its very high coordination costs and high compliance costs, smaller enterprises in the manufacturing sector are at a huge disadvantage. The distance to sophisticated buyers and the challenges with exports compounds the difficulty for smaller enterprises to compete globally from the local base.

Nelson is also known for his writing on the importance of a wide range of social institutions, both formal (for example a cluster development organisation) and informal (the trust networks between members of the clusters). He refers to these social institutions as social technologies, and he argues that they co-evolve with physical technologies to enable economic development. These social institutions range from central banks to a diverse range of firms, but importantly include other forms of organisations such as scientific and technological societies, universities, government agencies and even capital markets. These institutions are the focus of the discipline of innovation systems.

Nelson emphasises that “that when a potentially new technology emerges, new institutions often are needed to develop it, and invest in and operate effectively the economic practices based on it”.

Nelson acknowledges it is not an easy task, as it is hard to predict which emerging fields of promising new technologies are going to be important in driving economic progress in the future, and which will have a modest impact. The policies to create or reform institutions need to be adaptive and flexible. Arthur (2009:186) confirms the view of Nelson and argues that “We cannot tell in advance which phenomena will be discovered and converted into the basis of new technologies. Nor can we predict which combinations will be created.”

That brings me back to my intent with this post. When we look at technological disruption and change, it is very easy to get caught up in the potential or risks of any given technology. But we must not take our eye of the informal and formal institutions, market systems, regulations and technological domain specific organisations that are needed to make a new technology viable. At the same time, we also have to figure out how to gracefully exit older technologies and how to either shut down or transform public organisations that once had a critical role in supporting those industries and technologies.

Again, I repeat, the so-called fourth industrial revolution is going to be more disruptive at the level of institutions and social arrangements than it will be disruptive for the enterprises that are competing at the technological frontier.

In South Africa, we have a triple-challenge.

1 – Our institutions change very slowly, and we have huge social tensions about how to allocate resources and wealth in the economy. Our local municipalities and local economic development activities are ineffective (with some exceptions in some of the larger metros). Yet, local authorities have hardly any influence over the quality and effectiveness of national meso programmes that are supposed to enable economic change.

2 – This is compounded by a largely uncompetitive economy with lots of market concentration.  The regulatory burden in the economy keeps a lot of potential entrepreneurs employed in the corporate and the public sectors.

3 – Our discussions in South Africa about technological change, technological capability and the promotion of the innovation system is dominated by a linear logic of science leading to technology leading to innovation (the so-called STI approach). There is not enough attention being paid to the eco-system of organisations, technology extension agencies that can help enterprises master new technological domains, reduce coordination costs, the so-called Do, Use, Integrate (DUI) kind of innovation. On that point, we also have very few (if any) technological organisations tasked with transforming or upgrading whole sectors or regions in the country from a technological perspective. Everything is aimed at one enterprise at a time.

My research agenda:

This is what my research is about at the moment. I am working with a team from TIPS and the dti (South African Department of Trade and Industry) to strengthen the visibility of this technological meso network, while also strengthening the public sectors ability to spot technological disruptions and to be more pro-active.

Please sign up below if you want to stay informed of our progress as I will not be able to share all of our learning in the public space all the time.



Sources:

Arthur, W.B. 2009.  The nature of technology : what it is and how it evolves. New York: Free Press.

Nelson, R.R. 2015.  Understanding long-run economic development as an evolutionary process. Economia Politica,Vol. 32(1) pp. 11-29.

Saviotti, P.P. and Pyka, A. 2013.  The co-evolution of innovation, demand and growth. Economics of Innovation & New Technology, Vol. 225 pp. 461-482.

Using the S-Curve to identify potential disruptions

This is a continuation of my blog posts based on my research into how technological disruptions and change occurs.

A widely publicised model is the S-curve model that enables the evolution of the performance of a technology (Foster, 1986a; Foster, 1986b). In management of technology textbooks, this model is used to make predictions about the evolution of the rate of technological change, to detect possible technological disruptions, or to determine the limits of a particular technology.

In the S-curve model[1], the Y-Axis tracks the performance of a specific technology, while the X-Axis shows effort measured in R&D investment and resources aimed at technology development (see Figure 4).

s-curve

Figure 4: A technology S-curve

Source: Author, based on work by Foster and Christensen

In the beginning of the development cycle of a specific technology, it takes a lot of effort to get performance increases out of a technology (blue line in Figure 4). This phase is often characterised by many different competitors with many different approaches to solving a given technological problem. This is often followed by an exponential improvement curve where the effort pays off with large increases in performance. Typically, improvements to the performance of the technology at this point in time are driven by several incumbent competitors, with many companies if their technologies are not chosen. After a while, the performance increases for every unit of investment (effort) starts to taper and return on investment diminishes. This is where incumbent firms are most vulnerable, as they try to squeeze as much profit from their existing technologies without looking for new investment opportunities even though they are almost completely dominating the market. New entrants find it very difficult to challenge the incumbents in the existing market place, because the incumbents have established a brand reputation, distribution networks and supporting systems.

Clayton Christensen (2000) explains that incumbent firms are often overconfident about the value of their existing technologies and tend to ignore potential new technological approaches. New entrants that are using a different technology aimed at a different market segment might be entering a new steep S-curve (the red line in Figure 4).

The new technology is usually at a lower level of performance than the original technology, and targeting a small, not-so-profitable niche that the incumbent firms are not willing to fight for (as they are benefitting from the scale of their current customer base). The niche market provides the new technology space to innovate in ways to increase in performance, and at some point, the graphs may intersect. This is where whole industries or technologies can be disrupted, as existing customers switch to a new technology that is in an upward performance curve. Christensen and Raynor (2003) explain that incumbents are very often “relieved” to exit small, low-margin markets, and so they constantly upgrade towards higher-margin or higher-volume target markets. This leaves small niche markets for new entrants where demands are not being met. These niche buyers and the new entrants often work together through several development iterations together, until the performance curve of the new technology crosses the incumbent technology in the broader market.

Christensen argues that whether a technology is disruptive or not depends less on how radical it is, but more on its specific effect on the S-curve. If a new development improves performance of an existing technology, then the incumbents are preserved and tend to benefit most as this improvement often suits its current scale of operations. If a technology creates a new S-curve, then it may disrupt existing technology at some point, leading to a disruptive change in industry structure. This implies that radical or incremental performance improvements in most cases benefits incumbents, while disruptive innovation challenges industry structures. In an interesting twist, Christensen argues that incumbents are not ignorant of new technologies and underserved markets. He argues that they are the victims of their own success in making decisions that leverages existing knowledge, networks, markets and capabilities. Ironically, customers may actually communicate that they prefer incremental improvements on existing technologies rather than adjusting to disruptive technology. It is not only the innovator that faces risk and uncertainty, buyers also try to avoid making decisions about technologies that are only emerging, or where performance, results and requirements are vague or uncertain. Decision-making in research and development may also be biased towards the most likely-to-succeed ideas that directs resources away from tinkering or experimenting with fundamentally different ideas.

Existing companies may be able to spot an emerging technology or group of technologies with a potential to disrupt their current market. However, it may still be very difficult to decide when to switch more resources to completely new technologies that may also require different business structures, culture, market and supplier relations (thus switching resources from the blue line to the red line in Figure 4). The performance of the technology is born from the strategy of firms and how they allocate resources to product, process and business model innovation. One way that governments can reduce the costs of incumbents and new innovators to confront, investigate and test new technologies is through technology demonstration and applied technology research, where companies can visit, use or test technologies hosted by public universities. Because companies know that their competitors might be investigating the feasibility of trying a new technology, they themselves are more likely to invest in new skills, in trying the new technology or exploring how this new technology could result in new markets, business models and capabilities.

Gathering all the information that is necessary to construct an S-curve requires time and can be costly. It is especially difficult to figure out which performance criteria and measures of effort to use to construct the graph. However, when a portfolio of technologies is tracked this way it shows not only inflection points, but when certain technologies may outperform existing dominant technologies. A key question that must be answered in constructing this model is whether to track performance change at the level of components (modules), sub-systems or architectures. Furthermore, even if the performance lines cross, incumbents may not switch if their sunk investments are too high, or the learning cost of the new technology is too high. That is why newer companies are needed in the economy, as they might have lower sunk investments and more to gain from higher performance. Over time, resources shift from the old technology to the new, but only if the new technology is accepted and is disseminated sufficiently.

A critique of the S-curve model is that while the graph makes sense, it is often hard to construct and project into the future. It often makes sense ex-post to explain why a given technology outperformed a previous dominant technology. Also, a weakness of the narrow focus on technological performance disconnects the technology from the broader technological and social context, such as the organisation capacity and supporting networks and infrastructure that is required to make a given technology work.

Notes:

[1] It is called an S-curve because when the results are graphically illustrated the curve that is usually obtained is a sinusoidal line that resembles an S.

Sources

CHRISTENSEN, C.M. 2000.  The innovator’s dilemma: when new technologies cause great firms to fail. 1st Ed. New York, NY: HarperBusiness.

CHRISTENSEN, C.M. and RAYNOR, M.E. 2003.  The innovator’s solution: creating and sustaining successful growth. Boston, Mass.: Harvard Business School Press.

FOSTER, R. 1986a.  Innovation: the Attackers Advantage. New York: Summit Books.

FOSTER, R.N. 1986b.  Working The S-Curve: Assessing Technological Threats. Research Management, 294 17-20.

Citation for this text:

(TIPS, 2018:23-24)

TIPS. 2018. Framing the concepts that underpin discontinuous technological change, technological capability and absorptive capacity. Eds, Levin, Saul and Cunningham, Shawn.  1/4, Pretoria: Trade and Industry Policy Strategy (TIPS) and behalf of the Department of Trade and Industry, South Africa.   www.tips.org.za DOWNLOAD

Disruption, radical and incremental innovation

I am continuing with my sprint to write down the ideas and concepts that I am explaining regularly at different events and meetings (The earlier posts are all available on www.cunningham.org.za). The ideas about the fourth industrial revolution being disruptive are now discussed and repeated even by people who have very little understanding of technology or innovation, nevermind management (see post “what is the difference between 4IR and Industrie 4.0?”). There are subtle yet important differences between disruptive, radical and incremental innovation. These differences matter for policymakers, entrepreneurs and economic development practitioners.

In business management literature a distinction is often made between incremental, radical and disruptive innovation[1]. Incremental innovation introduces relatively minor changes to an existing product, process or technology, while radical innovation is based on a different set of engineering, scientific and business principles and often opens up new markets and applications. While incremental improvements may be small, the cumulative effects of an ongoing series of incremental improvements could be huge.

Incremental innovation exploits the potential of an established design and often reinforces the dominance of established firms. It mainly originates from within the sub-sector or system, and the informed or connected firms are often aware of the changing trends[2]. While it hardly requires new science, incremental innovation draws on incredible skills, deductive reasoning and experience, and over time can have significant cumulative economic consequences. Most businesspeople hardly recognise incremental improvement as innovation, although when prompted, many are able to identify several incremental improvements to their products, processes and organisational arrangements. Incremental innovations are chosen by the market if they offer savings, or add more value to what already exists. The market chooses an improved idea if it exceeds their existing expectations.

Radical innovations occur when new technologies are introduced into an existing market or technological domain. In the evolutionary technological change process, a radical innovation can start one of the change cycles (start a fluid phase), or it can be a blip in the performance of the technology during the amplification or selection phases.

Christensen (2000) argues that both incremental and radical innovations based on a specific technological paradigm often benefit incumbent firms, and describes them as sustainable innovations (for incumbent firms).  Incumbents and markets can recognise the benefits of the radical innovation and quickly adapt to it, or integrate it into their operations.

Disruptive innovation is different in that it often favours the new entrants (called the attackers by Christensen), who often combine different product, process and marketing innovation with a different business model. This part of the business model is really important.  Disruptive innovations are hardly about the product/service or the process, it is really about a different business model. These business model innovation often originate in niche markets where an innovator works very closely with niche market players to refine an idea in an iterative process before it is taken up by other markets. Incumbents have a really hard time to defend against this because they can copy some of the products, service or process features, but they often cannot copy the culture of the attacker.

Christensen et al. (2015) explains that disruptors often challenge incumbent firms with new business models, and attack incumbents by targeting marginal or even low-end markets[3]. Firms with resources and adaptive management systems are often able to exit these markets or to shift into new (often higher-value) market segments. While incumbents may be able to adapt their products and processes, it is often a matter of time before newer business models of the attackers outperform their traditional arrangements.

There are examples of famous and powerful firms going under or losing market dominance because they were disrupted by a new technological paradigm introduced by actors from another sector. Recent research comparing the US Fortune 500 companies in 1955 and 2017 shows that only 60 firms were in both lists (Perry, 2017). It is already hard enough for firms to stay abreast of technological changes and innovations within their sector and in related industries, therefore many established firms are often blindsided by technologies developed in other sectors that may in future disrupt them[4].

Some remarks about these ideas:

For most companies, radical and incremental innovations occur on a frequent basis. It may require rethinking a product, making changes to a process, finding new material suppliers or changing prices. While a competitor launching a new product, or announcing a change in pricing may disrupt your plans or cause a lot of stress, this is not what is meant with disruption. Disruption means that you cannot proceed in the same way. The markets you have served in the past now have new criterial which they use to select between alternatives and you have a weaker offer.

Disruptive innovations are disruptive because they require a rethink or demand a change of the core business model. Clients don’t want a price cut or an added feature. Some countries (like Singapore) promote disruptive technologies into their economies because it leads to increased innovation and much higher awareness by incumbents of global technological changes. Other countries try to defend against disruptive technologies, but in a way, they may only be postponing the inevitable.  What is clear to me is that companies cannot afford to only look for technological solutions within their industry or sector, but that they have to scan much broader. For an incumbent company to respond to a big disruption may require more business model innovation. For instance, our South African manufacturers have lost many competitive battles with manufacturers from Asia. Yet, very few manufacturers innovated in the business models by opening their own factories in Asia to learn from those markets.

Which brings me to a final remark. To get more companies hyper-sensitive to technological change, policymakers have to find ways to promote competition. It is only when small improvements make a big difference that incumbents would be willing to search beyond their current sectors for alternatives that offer even a small advantage.

Notes:

This is the 4th post that draws from the research and advisory work I am currently busy with to strengthen South Africa’s technological capability to detect and better respond to discontinuous technological change. The citation information for this post is at the bottom of this post, and a link to the research report that I have copied this from is here.

[1] While this literature is increasingly popular since the publications of Clayton Christensen, it is not new. Schumpeter (1934) and Freeman and Soete (1997), among others, already wrote about this much earlier.

[2] Several trends, such as the increasingly important knowledge-intensive business service sector, or new ways of sharing and protecting knowledge, play an important role in providing firms with access to new or relevant information.

[3] Christensen, Raynor and McDonald (2015) argue that from a disruptive theory perspective Uber is not seen as disruptive, as many taxis have been using apps for a long time, and Uber did not really enter the market by starting in underserved markets. However, due to the violent protests by traditional taxi owners, Uber is often described as being disruptive.

[4] An ironic example of a company that failed to recognise one of its own innovations as disruptive is Kodak. Management was so set on its film-based business and technology model that it chose to ignore its own market research that showed the disruptive potential of digital technology that one of their engineers developed in 1975. Not only did digital technology disrupt Kodak, it created many completely new industries, markets and applications.

 

Sources:

CHRISTENSEN, C.M. 2000.  The innovator’s dilemma: when new technologies cause great firms to fail. 1st Ed. New York, NY: HarperBusiness.

CHRISTENSEN, C.M., RAYNOR, M.E. and MCDONALD, R. 2015.  What Is Disruptive Innovation? Harvard Business Review, December 2015.

FREEMAN, C. and SOETE, L. 1997.  The Economics of Industrial Innovation. 3rd. London: Pinter.

PERRY, M. 2017. Fortune 500 firms 1955-v-2017.:   http://www.aei.org/

Citation for this text:

(TIPS, 2018:21-22)

TIPS. 2018. Framing the concepts that underpin discontinuous technological change, technological capability and absorptive capacity. Eds, Levin, Saul and Cunningham, Shawn.  1/4, Pretoria: Trade and Industry Policy Strategy (TIPS) and behalf of the Department of Trade and Industry, South Africa.   www.tips.org.za DOWNLOAD

Technological change cycles

This is the 3rd post that draws from the research and advisory work I am currently busy with to strengthen South Africa’s technological capability to detect and better respond to discontinuous technological change. The citation information for this post is at the bottom of this post, and a link to the research report that I have copied this from is here.

During the 1980s several scholars[1] recognised that technological change follows a cyclical pattern and several models were put forward to explain the phenomena. These models are still in use today and have been found to be active at different levels of technological change. The broad consensus was that a technological change cycle:

  1. Starts with a technological discontinuity or disruption, followed by a period of unstructured and often chaotic innovation when a new idea or concept is made possible (based on preceding developments). This disruption results in a fluid or turbulent development phase during which many ideas are developed, tried and promoted as the next best thing,
  2. That is followed by an era of ferment from which a dominant design emerges; and
  3. This is followed by an era of incremental change during which the dominant design is elaborated.

This can be illustrated with the widely recognised Abernathy and Utterback (1978) model with its three phases of change that are illustrated in Figure 2. The three phases are a fluid phase, a transitional phase, and a specific phase, and is similar to the cyclical pattern described in the bullet list above. Other scholars used slightly different labels, but the characteristics in the different phases are all more or less the same.

Abernathy and Utterback

Figure 2: The Abernathy-Utterback model of technological change

Source: Abernathy and Utterback (1978)

The rate of innovation is highest during the fluid phase, during which a great deal of experimentation with product features and operational characteristics takes place between different competitors[2]. Because of all the changes in the product composition and characteristics, process innovation typically lags. Buyers and users are often confused or overwhelmed during this phase fearing that the benefits are overstated and that the costs of adaptation are uncertain. Only the brave and the innovative engage in finding, adapting and integrating new ideas and concepts.

In the transitional phase, the rate of product innovation slows down and the rate of process innovation increases. At this point, product variety gives way to standard designs that have either proven themselves in the market, or that are shaped by regulations, standards or legal constraints. The pace of innovation of how to produce the product increases. What was done earlier by highly skilled technicians may become automated or developed to a point when low-skilled operators can take over. Or lower-skills jobs are displaced from the production process to other functions like logistics, while the skills intensity on the production line is enhanced. At this point it is easier for bystanders and followers to engage in exploration. The early adopters are already over the horizon, while many early adopters have exited, sold out or moved on.

The final phase, the specific phase, is when the rate of major innovation dwindles for both product and process innovation. In this phase, the focus is on cost, volume, and capacity. Most innovations are very small incremental steps, improvements on what is already known and accepted. Latecomers can now engage with the technology, although it might already be too late.

The description of technological change provided above follows the generic three-step process of technology evolution: a process of variety creation, selection, and then amplification or retention.

  • During the variety creation phase there are many competing designs and no dominant logic. Towards the end of this phase a few dominant designs may emerge, but there is still much competition between ideas. This is not only a technical selection process, there are important social, political and industrial adjustments taking place at the same time.
  • During the selection phase, standards emerge for positively selected ideas, with a few designs dominating. It is a relatively stable process of incremental improvements in features, performance and results. This may be interrupted occasionally by leaps in performance as some designs are substituted by better technologies, or from breakthroughs often coming from other industries or contexts. In general, designs become simpler as a learning process unfolds about how best to design, manufacture, distribute and use a particular technology around dominant designs. This period is characterised by growing interdependence as modules are developed, substituted and standardised. There is a growing exchange and increased competence within and between different communities of practitioners. Often there is industry consolidation during this phase. It is important to note the dominant designs are only visible in retrospect. They reduce variation, and in turn, uncertainty, but within the process it is hard to predict which designs will survive the next set of radical innovations. Once a design becomes an industry standard it becomes hard to dislodge.
  • This leads to an amplification phase, in which the best ideas are not necessarily used as intended, but when technological changes spill over into areas not originally intended. This is a relatively stable process that can continue for long periods, until is it suddenly interrupted by a radically different idea, resulting in the process starting all over again.

Anderson and Tushman (1990) state that, from the perspective of the sociology of technology, technological change can be modelled as evolving through long periods of incremental change punctuated by revolutionary breakthroughs[3]. The innovation activities that take place that lead to these phenomena will be discussed in Chapter 3.

Arthur (2009:163) contends that change within technological domains is a slow process. He explains technology domains do not develop like individual technologies like a jet engine: focused, concentrated and rational. It is rather more like the development of legal codes: slow, organic and cumulative. With technology domains, what comes into being is not a new device or method, but a new vocabulary for expression, similar to a new language for creating and combining new functionalities.

A current example is the “Internet of things”, where the connectivity of physical devices are spreading from the office and smartphone devices to interconnect household appliances, industrial applications and an endless list of technologies enabling data exchange, control and new functionalities . It could be argued that this is not a new technology, digital sensors have been around for a long time, our cars, smartphones and equipment have contained them for a long time. However, the language, standards, distributed nature of processing, and developments in big data visualisation have all contributed to this technology appearing to arise from obscurity into the limelight of the popular media. A similar argument could be made for artificial intelligence, drone technology and others.

Notes:

[1] The work of Tushman and Anderson (1986), Abernathy and Utterback (1978) are still frequently cited today.

[2] Kuhn (1962) noted that in the early stages of research in a given field, the most that scholars typically can do is to report the phenomena they observe, without a unifying theory or framework to help them categorise or make sense of what they see. As a result, this stage of knowledge accumulation is characterised by confusion and contradiction. Theories are put forward but reports of deviating phenomena accumulate.

[3] This is often referred to as punctuated equilibrium by political scientists.

 

Sources

Abernathy, W.J. and Utterback, J.M. 1978.  Patterns of Industrial Innovation. Technology Review, Vol. 80No. 7 (June/July 1978) pp. 40-47.

Anderson, P. and Tushman, M.L. 1990.  Technological Discontinuities and Dominant Designs: A Cyclical Model of Technological Change. Administrative Science Quarterly, Vol. 35No. 4 (Dec 1990) pp. 604-633.

Arthur, W.B. 2009.  The nature of technology : what it is and how it evolves. New York: Free Press.

Kuhn, T.S. 1962.  The Structure of Scientific Revolutions. Chicago & London: University of Chicago Press.

Tushman, M.L. and Anderson, P. 1986.  Technological Discontinuities and Organizational Environments. Administrative Science Quarterly, Vol. 31No. 3 pp. 439-465.

 

Citation for this text:

(TIPS, 2018:12-13)

TIPS. 2018. Framing the concepts that underpin discontinuous technological change, technological capability and absorptive capacity. Eds, Levin, Saul and Cunningham, Shawn.  1/4, Pretoria: Trade and Industry Policy Strategy (TIPS) and behalf of the Department of Trade and Industry, South Africa.   www.tips.org.za DOWNLOAD

 

 

Technological architectures

An important distinction can be made between architectural innovation and component-level innovation. The architecture defines the way different components or subsystems are organised and how they interact with other components. Often architectures themselves form part of even larger webs of architectures.

Innovations at the component level, which is a physically distinct portion of the technology that embodies a separate design concept, mostly reduce costs of production, and often take place at high frequency with a wide range of choices available. While the organisations that innovate at the component level are more dependent on past experience as well as economies of scale, the organisations that determine the architecture are able to depend far more on their value addition, as well as the sunken investments of many other agents into the system.

To change the architecture of a system requires many simultaneous changes to different sub-architecture and component levels, which may be beneficial to some agents in the system, but not to others (thus vested interests often create a path dependency). A change to the architecture could even disrupt industry structure, and it changes the way the markets judge whether a specific architecture is suitable for the function or tasks it fulfils. A combination of path dependency and architectural change can be used to describe why many industries (or architectures) have disappeared.

However, architectures such as the vehicle example in the figure above change slowly over time and can certainly be influenced by improvements at the component level. For instance, better electronic management of the engine may result in less frequent services, but the architecture hardly changes. Interestingly, the architecture of the vehicle also forms part of a wider architecture of road networks and urban designs, again reinforcing another higher level of path dependency. This nested nature of technologies at the level of architectures is what slows down massive technological change. To continue with the example of a car, passenger vehicles depend on the architecture of a road network. It is also dependent on fuel and maintenance systems, parking arrangements, insurance and all kinds of traffic and safety laws.

I find it interesting that two decades ago, electric vehicles were described as being massively disruptive resulting in the demise of the fossil-fuel vehicle. Now, many established car manufacturers have jumped onto the bandwagon and are investing heavily in their own electric vehicle technologies, and in doing so reducing the disruptive effect of alternative fuels. In doing so, they are making massive strides in fuel efficiency, reducing the weight of their cars and substituting harmful and heavy materials with materials that have less impact on the environment. The component and sub-system level innovations offered by electric vehicles are being incorporated into the designs of the older fossil fuel architecture, while the architecture itself is only changing slowly. In South Africa, the network of charging stations and points are slowly expanding, but the use of electric vehicles is still minute compared to the fossil-car usage.

Some examples of architectures and components are computers (architecture) and an internal graphics card (component) or a jet airliner (architecture) and in-air entertainment systems (components).

The reason why I thought it a is a good idea to go back to such a basic distinction as the difference between architectural innovation and component level innovation is that in much of the popular discussion about technological disruption (the fourth industrial revolution-talk) this distinction is not made. What I appreciate about the World Economic Forum is that they are raising awareness of what will happen to social arrangements when one architecture displaces another. But what is not receiving enough attention are the many challenges that we will face in developing countries at the level of sub-systems and components. This is where competitiveness, resilience and innovation are critical because this where the disruptions and discontinuities of industries will occur. This is also the area where developing countries usually follow (as outsourced manufacturers) and where we are the most vulnerable to the design capabilities and dense networks that existing in clusters in the developed world.

I will explore how these changes occur in the next few posts.

Is the Fourth Industrial Revolution a paradigm shift?

I am excited that the Helvetas Eastern European team asked me to write a blog post for their Mosaic newsletter about the Fourth Industrial Revolution. The blog article and many others can be found here.

Regular readers will know that I am not so convinced of one big revolution; rather that there are many smaller disruptions. In this article, I argue that it is hard to imagine what a paradigm shift would look like. I make six arguments of why there are rather several smaller disruptions taking place. The credit for coming up with the image in the article goes to Zenebe Uraguchi from Helvetas. He is also the person that convinced me to write this article, and who guided me when I felt stuck. Thank you, Zenebe! Take a look at some of Zenebe’s posts on the Inclusive Systems blog of Helvetas.

The second half of the article I wrote is about figuring out which social technologies to develop that are needed to make certain technologies usable or beneficial to societies. Many of these social technologies are cultural or organisational, but there are also many public institutions and public goods that are lacking in developing countries.

To me, it feels that we are still just scratching the surface when it comes to helping the meso organisations of developing countries cope with technological change.

However, it is exciting that my research into discontinuous technological change and the necessary social and technological institutions that are required in developing countries is of interest to development organisations and governments.

I am looking forward to your comments, questions, contradictions and ideas!

Best wishes,

Shawn

Shawn in Wonderland

I have not posted anything for the last 3 months. I have been on an amazing adventure which is so similar to Alice in Wonderland that I might be asleep and still dreaming.

It started with a long-pursued opportunity to help a unit in the South African government prepare and think through the consequences of the “fourth industrial revolution” and the fuzzy collection of Industry 4.0 gadgetry that will soon overthrow our lives. By all popular accounts, this revolution will smack us hard, because the narrative in South Africa is that we are behind and falling further behind. The prophets blame all our usual reasons for this impending doom: our poor education system, our unskilled workforce, an unemployable youth, labour unions, capitalist greed, our government policies, inequality, high costs of everything, low public investment, corruption and the easter bunny. (OK, I made up the last one.)

Now don’t get me wrong. I know we are drifting sideways in many respects, maybe even regressing in some areas. For example, our economic complexity is in decline. Our technological capability is dropping. Many of our traditional sectors are uncompetitive. I have been working in the high-tech sectors and I know how hard it is to get to any kind of scale. Our institutions struggle to adapt, are underfunded, and our business people face high uncertainty, as much uncertainty as our public officials.

It is clear to me that the pace and convergence of change is increasing. The amount of information is increasing. We all are drowning in documents, reports, blog posts, emails, journals and correspondence. The demands both on specialists and generalists are increasing. So there is definitely something cooking.

But is it an industrial revolution?

Are revolutions not full of social unrest, upheaval of institutions, overthrowing of  government structures?

That is the big question that I started with. I must admit the empirical and academic evidence is thin on this topic. The only people excited are geeks and suppliers of gadgets. This really bothered me, so I tried to figure out what all the things are that I would have to understand to sense, monitor, track and possibly predict where technologies are changing, how these shifts could affect our institutional structures, industries and jobs.

So I went on the most amazing reading journey. Down the rabbit hole I went.

I started by exploring the literature on how technological change happens, how technology cycles unfold. I could get lost in little forests of papers, books and articles by many of my favorite scholars. I followed ideas down paths (to the 1980’s) and came back again to 2018. Actually, not much has changed since the early writings of Nelson, Pavitt, Lall, Freeman, Edquist, Perez and many others. I admire these scholars because they really grasped the principles at such a fundamental level that not even the arrival of the internet really nullified any of their theories. I then investigated technological evolution and was again inspired by the clear writing of Arthur, Hidalgo, Hausmann and Rodrik (on structural change and industrialisation).

Then I stood back and wondered about all the innovation, tinkering, risk taking and failing that had to happen to lead to the patterns that I found in the chapter on technology. Again, I went into a forest, this time looking at innovation, how it happened, did not happen and why. I was inspired by the work of Dosi, Fagerberg, Malerba, Dodgson, Teece, Utterback, Clark, Henderson and Christensen.

For a week I felt paralysed by these two forests. Are they really two different domains deserving separate chapters, or should they be integrated into one? In the past I have treated them as separate. So, I procrastinated and forged into one of my favorite topics, that of innovation systems and how they change.

It was always my intention to hold back on this walk into the innovation system forest, as I wanted to look at everything here with new eyes. I plunged into my favourite authors, Nelson, Dosi, Freeman, Fagerberg, Srholec, Lundvall, and some more Nelson, and many other authors I admire. I was again struck by the importance of building technological capability, increasing absorption capacity and the importance of social, technical and other meso organisations in all of this.

Towards the end of the innovation system week I ventured into the work of Johan Schot and Frank Geels, Andy Sterling and Ed Steinmuller (the SPRU network), and got lost in the world of socio-technical transformation. I could look at the literature on institutional change and discovered the work of Thelen. I spent a whole day just reading up on Carlota Perez, and the next day I went back to the earlier works of Christopher Freeman (which then lead me down the archives of the SPRU). Perez is one of the few scholars who even mention the word “revolution” and she argues that developing countries must embrace rapid technological change to achieve structural change.

I came out of this forest dazed, confused and inspired. All at the same time. I decided I had to integrate my innovation chapter into the technology chapter. It took me three days to integrate them. I also tried to integrate the socio-technical transformation section into innovation systems.

Then I went away on a weekend in the Bushveld in the Limpopo province in South Africa. Somewhere while breathing fresh air in the country-side I realised that technology and innovation had to be separated, largely because there is a tendency in South Africa to focus on linear innovation (science=>technology application => innovation). I recalled something that my late business partner and friend Jorg Meyer-Stamer repeatedly said.

“Technology is about action, about harnessing natural phenomena to achieve something. Innovation is about a difference, it is about doing something differently”.

For my client to measure and track technological change would not be too difficult. Measuring innovation will be much harder, as a lot of the innovation caused by the “revolution” are about changes in social technologies, organisational culture and strategy.

Four weeks into my study and I was left with one messy section. It involved reconciling my views on innovation systems with the socio-technical transformation and multiple pathways literature. It felt like I was stuck in mud. The common factor between these fields is the importance of adaptive meso institutions, tied with a balanced supply side and demand side interventions. Context matters in both these fields, far more than firm level technological use and innovation practices. What I like about the social technical transformation literature is their focus on developing “niches” based on unique contextual opportunities or challenges, and their recognition of how change unfolds and spills over in time. Too often innovation systems treats the system like a static network of publicly funded organisations.

So that is where I am now. My first draft literature study is complete. I’ve had so much fun during this journey. You would notice that I did not mention economic complexity much. The days that I somehow cannot account for was spent on that, but I really tried not to get sucked in too deep. In the end I decided not to include this in this study.

Stay tuned for a future update about what I discovered.