Chapter 1. Understanding digital transformation

The statistical data for Israel are supplied by and under the responsibility of the relevant Israeli authorities. The use of such data by the OECD is without prejudice to the status of the Golan Heights, East Jerusalem and Israeli settlements in the West Bank under the terms of international law.

Introduction

Reaping the benefits and addressing the challenges of the digital age requires narrowing the gap between technological developments and public policies. Many policies are the legacy of the pre-digital era, and difficulties in understanding the changes underway and their implications may delay the review and adaptation of these policies. Such an understanding is imperative as digital transformation affects the entire economy and society.

1.1. What is digital transformation?

Digitisation is the conversion of analogue data and processes into a machine-readable format. Digitalisation is the use of digital technologies and data as well as interconnection that results in new or changes to existing activities. Digital transformation refers to the economic and societal effects of digitisation and digitalisation.

To develop policies fit for the digital age, it is critical to be aware of the main elements of the evolving digital technology ecosystem and some of the opportunities (and challenges) resulting from their application. Second, it is essential to understand the data revolution that is taking place, and how data and data flows affect individuals, the economy and society more broadly. Third, it is important to identify the key properties of digital transformation, including how they are driving new and evolving business models, and what their implications are for public policy.

The digital technology ecosystem

Dramatic increases in computing power and a simultaneous decline in related costs over the last 60 years has driven the rapid advance of digital technologies (OECD, 2015_[1]; Moore, 1965_[2]). Today, an ecosystem of interdependent digital technologies underpins digital transformation and will evolve to drive future economic and societal changes (Figure 1.1).

This ecosystem is much stronger and functional than its individual components because they interoperate with and complement one another, opening up new possibilities. Some of these technologies have already arrived and are part of our daily lives. Others are still on the horizon. All of these technologies hold potential benefits for growth and well-being.

Internet of Things

The Internet of Things (IoT) enables a host of new business models, applications and services based on data collected from devices and objects, including those that sense and interface with the physical world. IoT devices involve those with both short- and long-range communication connectivity. Massive machine-to-machine (M2M) communications which are composed of sensors for smart cities, agriculture, manufacturing and the like are a subset of the IoT.

The IoT includes automations from smart home devices and appliances, wearables and health monitors, to advanced applications like connected and autonomous vehicles. In fact, today airplane turbines constantly gather data and can relay them when a problem arises. When the plane lands, a maintenance crew is ready with the right parts and knowledge of what the issue is, alleviating delays. Remote mining and surgery, for example, will also become practical thanks to the IoT. Utilities will be interconnected with millions of networked devices, allowing them to take more informed decisions autonomously and in real time. Moreover, Internet-connected sensors and actuators will monitor the health, location and activities of people and animals, the state of the environment, and much more (OECD, 2016_[3]).

Next-generation wireless networks: “5G” and beyond

While the international standard is not yet finalised, 5G will be the first generation of wireless networks conceived mainly for a future in which tens of billions of devices and sensors are connected to the Internet.1 Major improvements upon previous network generations include higher speeds (i.e. 200 times faster than 4G), faster data transfer (i.e. 10 times less than 4G), and networks that better support diverse applications through the virtualisation of the physical layers (i.e. “network slicing”). Trials are underway in multiple countries, including through collaborations between network operators and vertical industries such as the automotive industry (OECD, forthcoming_[4]).

A major difference with 5G is that it is designed to connect not just people, but things, underpinning a world of M2M communication that takes place largely hidden from human eyes. 5G networks will improve communication between self-driving vehicles, roads and traffic lights, making “platooning” feasible – the automatic linking of vehicles on highways in a convoy so that they are much closer together than what would be safe with human drivers. This could ease road congestion as well as improve safety and fuel efficiency. In addition, sensors embedded throughout farms will be able to communicate crops’ water and fertilisation needs directly to agricultural machinery and systems. Personal devices will download data at far higher speeds even in crowded areas, realising the potential coverage of on-demand media from almost any location reached by 5G networks.

Cloud computing

Cloud computing is a service model that provides clients with flexible, on-demand access to a range of computing resources (OECD, 2014_[5]). Clients access such resources (e.g. software applications, storage capacity, networking and computing power) online. The resources can be used (and priced) in a scalable and adaptable manner, enabling customers to transform substantial fixed costs for information and communication technologies (ICTs) into lower marginal costs, and to more easily match their supply of ICTs with their evolving business needs. In other words, cloud computing allows users to rent the ICTs they need at any given time rather than having to buy them outright. Cloud computing increases the affordability, availability, capacity, variety and ubiquity of computing resources in a way that facilitates other digital technologies, such as artificial intelligence (AI), autonomous machines, big data and 3D printing, as well as the wider digital transformation (OECD, 2015_[1]; OECD, 2017_[6]).

Cloud computing applications abound and go well beyond simple storage of personal files, photos and videos; they also allow remote access and enable people to collaborate on documents at a distance. For example, personal CD and DVD collections are becoming things of the past as we migrate to streaming audio and video services like Deezer, YouTube and Netflix, all of which are made possible by cloud computing. We can carry access to our entire personal libraries around with us on a single tablet, thanks to ebooks that are stored in the cloud. Duplicating copies of computer files onto a locally connected hard drive and then manually moved off-site for disaster recovery purposes is no longer needed. Instead, backup and disaster recovery services are directly accessible in the cloud. Mobile apps reside in the cloud and often rely on it to function even after they are downloaded. Smart thermostat systems use cloud computing to monitor, analyse and adjust to temperature trends in homes, leading to less energy consumption, lower utility bills and “greener” living.

Big data analytics

The term “big data” commonly refers to data characterised by high volume, velocity and variety. It benefits from the IoT, among other technologies, as a source of data and from cloud computing as a source of processing power. While large quantities of data can have value in themselves, namely when commercialised, most of their value depends on the capacity to extract information from the data. Big data analytics techniques and software tools are used, for example, for data (text) mining, profiling and machine learning. By fostering new products, processes, organisational methods and markets, and improving existing ones, the use of (big) data analytics enables data-driven innovation and the potential to improve productivity and well-being (OECD, 2015_[1]).

Big data analytics have enormous potential, some of which has already been realised. For example, retailers routinely use big data analytics to make tailored suggestions to customers based on the customers’ interests as revealed by their prior browsing and shopping behaviour. In a quite different setting, neonatal units monitor the heartbeats and breathing patterns of premature and sick babies, feed the data into an ever-growing database and, with the assistance of analytics, can predict infections 24 hours before the babies show any physical symptoms.

With enough data from developing countries, governments and aid organisations can maximise their impact by using big data analytics to identify areas where people will benefit the most from better access to education, healthcare and infrastructure. Epidemiologists can take big data from search engines into account when finding and tracking contagious disease outbreaks. Competition authorities can fight corrupt business practices such as bid rigging more effectively with the help of big data, which can be used to identify suspicious bidding patterns. Physicists also benefit from big data, which has made projects such as the Conseil européen pour la recherché nucléaire’s (CERN’s) Large Hadron Collider (LHC) possible. The LHC produces 30 petabytes2 of data per year. The CERN’s data centre contains 65 000 processors, but it also uses thousands of other computers across 170 other data centres to analyse its data. Big data also underlies AI.

Artificial intelligence

AI is the ability of machines and systems to acquire and apply knowledge, including by performing a broad variety of cognitive tasks, e.g. sensing, processing language, pattern recognition, learning, and making decisions and predictions. Much recent progress in applying AI is driven by machine learning (when machines make decisions based on probability functions derived from past experiences), big data analytics, dramatically increased processing power and cloud computing, all of which enable AI to process data at enormous scales and to accelerate the discovery of patterns in data. AI drives new kinds of software and robots that are increasingly: 1) “autonomous” or semi-autonomous, meaning they make and execute decisions with no or little human input; and 2) capable of learning, evolving and improving throughout their life cycle to customise and improve functionality and performance based on the analysis of data collected from their environment.

AI is already part of daily life in many countries. Learning algorithms detect patterns in our digital behaviour and use them to influence the search results and advertisements we see, the news we read and the entertainment we consume. For instance, recommendations on Amazon, Netflix and Spotify are based on machine learning technologies. AI helps doctors to detect, track and treat diseases. Robotic surgeons are already in use. Furthermore, algorithms now conduct more stock market trades autonomously than humans in the United States (OECD, 2015_[1]). AI applications still hold many promises for the future as well. For example, AI will eventually give robots the ability to adapt to new working environments with no need for reprogramming.

Someday, AI-powered robots might care for elderly people, tending to their physical needs while interacting with them. In the future, AI might sift through databases of medical histories to develop tailored treatment plans that are most likely to work for individuals with a given set of characteristics, replacing one-size-fits-all approaches. Some people find some of AI’s future prospects unsettling, though, such as its use in driverless cars or in robots that could displace significant portions of today’s workforce.

Blockchain

Blockchain is a technology that enables applications to authenticate ownership and carry out secure transactions for a variety of asset types. It is a ledger or a spreadsheet that is maintained and stored across a network of computers. The network regularly updates the database in every place it exists, so that all copies are always identical. This means the records are visible and verifiable to everyone else in the network and there is no need for intermediaries to serve as authenticators. New events and transactions are automatically stored in “blocks” which are then chained to one another chronologically using advanced cryptography, creating a digital record. Should someone try to change information stored in the block, the “chain” is broken and all nodes in the network are aware of it. It is for this reason that this technology is called blockchain and that it is often described as tamper resistant.

Blockchains can be public, called “unpermissoned”, whereby access and transfer occur between parties unknown to one another (e.g. Bitcoin). In contrast, private or “permissioned” blockchains enable access and transfer between specific parties, and are executed much more quickly. Some blockchains can also execute software in a decentralised manner, without the need for intervention from a central operator. This means that some applications, often known as “smart contracts”, can execute in a pre-defined and strictly deterministic manner. The third generation of blockchain technology, which is currently unfolding, allows interoperability across different blockchains.

One of blockchain’s most widespread application so far has been for cryptocurrencies (e.g. Bitcoin, Ripple), but it is starting to affect many other sectors, including agriculture, manufacturing, retail, healthcare, energy, transport and the public sector. Eventually, a major application of blockchain may be securing data in the cloud. It could also be used to make everything from charitable donations to elections more verifiable and secure. However, the immutability of blockchain may impact the “right to be forgotten” in some jurisdictions.

Computing power

High-performance computing (HPC) is the aggregation of processing power to deliver far higher performance than would be possible with an ordinary computer. HPC is typically used to solve big science, engineering or business problems. It can also be used for other purposes, such as in the well-known case of DeepMind’s AlphaZero, to run software that trains itself how to play board games. In fact, it took only nine hours of training for AlphaZero to defeat world champion Go and chess programmes. HPC is growing more important for firms in a wide variety of industries, including construction, pharmaceuticals, the automotive sector and aerospace. The ways in which HPC is used in manufacturing are also growing, as they now include not only applications such as design and simulation but also real-time control of complex production processes.

Quantum computing (QC) takes a fundamentally different approach. Traditional computing processes data in an exclusive binary state at any point in time (that is, bits take a value of either 0 or 1 and cannot be a superposition of 0 and 1). In contrast, QC relies on “qubits” which are organised in “states” that represent some combination of 0 and 1 (Metodi, Faruque and Chong, 2011_[7]). The qubits, even when separated by huge distances, can interact with each other instantaneously (they are not limited to the speed of light). “Entangled” with each other in pairs through a process known as correlation, they can be used together with an algorithm to answer questions. This is an emerging field and substantial obstacles still need to be overcome. For example, most of today’s experimental quantum devices must operate in temperatures near absolute zero and require the development of new materials. However, if it succeeds, QC would be an enormous leap in processing power due to its ability to operate in multiple states and to perform tasks using all possible permutations simultaneously.

QC’s ability to process information at almost unthinkably fast speeds versus today’s ICTs would make it perfect for AI and cloud computing. This is because they require network systems that do not get bogged down with heavy use. In addition, if blockchain secures much of what is stored in the cloud, QC would become even more useful in light of the formidable computing power and electricity required to complete blockchain transactions. Quantum computers could also be used in simulators that replicate real physical systems, allowing manufacturers to design things such as better batteries and satellites or new materials for airplanes. While QC may challenge existing digital security technologies such as cryptography, it could also be used to support new ones.

The combination of technologies in one digital ecosystem multiplies their potential

Each technology alone can bring its own opportunities and challenges, but the biggest potential lies in their combination within one digital technology ecosystem. For example, cloud computing’s effectiveness requires always-on, everywhere-available and high-speed Internet connectivity and is essential to big data analytics, which also relies on powerful computing. The use of billions of devices and sensors in the IoT generates big data that are a key resource for sophisticated algorithms and machine learning, enabling AI to be used in an ever-growing range of areas, and turning AI itself into a resource.

Thanks to a confluence of technologies, machines can view and understand images and videos (“computer vision”). Consequently, a machine in the cloud using AI can communicate with drones over 5G networks, enabling them to identify anything from license plates on a vehicle to a leak in a pipeline in real time. Finally, the smartphone illustrates how the use of many key digital technologies, e.g. fast connectivity, access to cloud services, multiple sensors, AI, etc., have already become ubiquitous and play an increasingly important role in everyday life. Assessing the opportunities and challenges created by the use of each of these technologies alone and in combination is thus essential to developing policies well-suited to the digital age.

The data revolution

The digital technology ecosystem relies on data. Data increasingly underpin digital transformation and have become an important source of value, for example for decision-making and production. While issues around data span across policy areas and are addressed throughout the report, it is important to first understand data as a critical resource and a source of value, as well as some transversal policy challenges related to data.

Recognising data as a critical resource

Data have been collected ever since humans recorded facts as symbols such as numbers, but the volume of data collected in the past is a drop in today’s growing data ocean. Every day, more data are produced than since the dawn of civilisation up until the early 2000s (Siegler, 2010_[8]): roughly 5 Exabytes, which corresponds to 1.25 billion DVDs (CISCO, 2017_[9]). Until recently, humans recorded most data themselves and often on rigid materials such as paper. Today, most data are collected by machines that are equipped with great storage capacity, powered by fast processors, and connected to the Internet.

Key technologies that produce and use data have become so ubiquitous, small and inexpensive that over a third of the global population carries a smartphone. In turn, connected devices, and smartphones in particular, are central platforms for data collection and consumption, alongside the IoT with its growing number of sensors and actuators embedded in devices, infrastructures and environments.

While data sources are multiplying, a majority of the data exchanged over global Internet protocol (IP) networks, notably the Internet, is created and used by consumers, in particular Internet videos. In 2018, Internet videos represented 49% of global IP traffic and 76% of global consumer Internet traffic; by 2022 the respective shares are projected to reach 61% and 82%. Meanwhile, the fastest growth in Internet traffic is expected to occur on mobile networks, driven by 47% annual growth (compound annual growth rate) of mobile consumer Internet traffic between 2017 and 2022 (Figure 1.2).

Data have become an important and valuable resource. Data are not a natural resource like oil, water, or air: they are created by humans and produced through human (and increasingly machine) activity. Data can be characterised as being of general purpose, non-rivalrous, and a capital good3 (OECD, 2015_[1]). In contrast to natural resources, the volume of data increases with its collection and use. Digital data can be copied and re-used endlessly, enable economies of scale and scope, power AI, and be used to improve existing or invent new products and (virtual) reality. This also means that the economy’s function of allocating scarce resources may be affected would data as an abundant resource be accessible by all.

Digital data differ from analogue insofar that they can be used, re-used, copied, moved, and processed cheaply, without degradation, and very fast. Again, in contrast to natural resources, processing and movement of data are neither constrained by gravity nor by material resistance. Data can flow at the speed of light, underpinning its velocity, between people, businesses and machines, across borders and the globe in milliseconds, thanks to the first truly global infrastructure ever built, the Internet. Delivering exploding volumes of data around the globe, the Internet has evolved as a network of networks, made up of cables, exchange points, masts, etc.; however, fast data delivery increasingly relies on local caching of data, close to where people demand and expect it, at their fingertips (Box 1.2).

1.2. Content delivery networks and local caching of data

Content delivery networks (CDNs) serve as aggregators of content, systems for the delivery of traffic directly to the terminating network, and providers of quality-enhancing inputs, such as caching of data close to the end user. CDNs are useful to providers of online services, such as the BBC, Google, Netflix and Hulu, which seek to improve their customers’ experience. More direct delivery, fewer intermediate loops, and local caching reduce latency and improve the quality of service.

Local caching of data reduces the volume of traffic that needs to be delivered to the terminating network. Caching refers to the storage of data locally so that data requests can be referenced to previous results and responded to faster. This means that it isn’t necessary to access the same data, potentially far away, in response to a similar previous request. As such, CDNs are used by YouTube for fast delivery of high-quality videos via local caches close to the consumer.

Source: Weller and Woodcock (2013_[11]), “Internet traffic exchange: Market developments and policy challenges”, https://doi.org/10.1787/5k918gpt130q-en.

Data are not homogenous. In theory, the variety of different types of data is infinite. In practice, many approaches are being developed to distinguish between different types of data and data flows, including by the OECD (Box 1.3). Other distinctions include, for example: public sector versus private sector data; personal versus non-personal data (Hofheinz and Osimo, 2017_[12]); user-created versus machine-generated data; data distinguished by the actors exchanging it, such as business to business (e.g. financial or IoT), business to consumer (e.g. media, consumer), government to user (e.g. services), or consumer to consumer (e.g. communications, social) (Kommerskollegium, 2014_[13]); qualitative versus quantitative data; structured versus unstructured data; or data distinguished by their origin, e.g. whether data is provided, observed, derived, or inferred, etc. (OECD, forthcoming_[14]).

1.3. Disentangling different types of data

Among multiple ways to possibly break down data into different types, one approach developed by the OECD with relevance for policy making distinguishes the following:

Personal data include data that allow for the identification of an individual data subject (OECD, 2013_[15]). They can cover public and private sector data, e.g. user-generated content (i.e. blogs, photos, tweets) or geo-location data from mobile devices as well as public sector data (i.e. police records, social security numbers).
Public sector (government) data include data that are generated, created, collected, processed, preserved, maintained, disseminated, or funded by or for the government or public institutions and include open government data.
Private sector data complement public sector data, namely as data that is generated, created, collected, processed, preserved, maintained, disseminated and funded only by private sector.
Proprietary (private) data include public or private sector data protected by intellectual property rights (IPRs) (e.g. copyright and trade secrets) or by other rights with similar effects (e.g. provided by contract or cyber-criminal law).
Research data include factual records (numerical scores, textual records, images and sounds) used as primary sources for scientific research, and that are commonly accepted in the scientific community as necessary to validate research findings.
Public (domain) data are not protected by IPRs (or other similar legal rights) and therefore lie in the “public domain” and are publicly available, free for use by anyone for any purpose without any legal restrictions.
Data of public interest include public or private sector as well as personal or non-personal data needed to fulfil well-defined societal objectives that otherwise would be impossible or too costly to fulfil.

Source: OECD (forthcoming_[14]), Enhanced Access to Data: Reconciling Risks and Benefits of Data Sharing and Re-use.

Extracting insights from data creates value

Data by themselves do not necessarily have intrinsic value. Their value not only depends on volume, variety and velocity of data (i.e. “big data”), but also depends on their veracity, quality or fitness for use, and other factors inherent to the data (OECD, 2011_[16]). Specific characteristics of data may be more valuable for some users than for others, e.g. velocity is crucial for an application providing traffic updates, but much less so for an online genealogy service. This illustrates that the value of data depends on the context and the potential benefit of their use. More specifically, data become valuable when information can be derived from them, and such information is always context-dependent (OECD, 2013_[17]).

Data analytics are essential to extract insights from data and to create value. Data analytics include a set of techniques, tools – software, AI, visualisation tools, etc. – that help extract information from data by revealing the context in which the data are embedded and their organisation and structure. Effectively analysing data with such tools crucially requires human capacity, notably skills, such as data analytic and management skills. Data analytics help extract information from data, which can be used to generate knowledge and/or support decision making. Rather than being a linear process, value creation from data takes place in a value cycle with feedback loops at several stages of value creation. This value cycle reflects an ongoing process of: datafication and data collection, structuring big data, extracting insights through analytics, constituting a knowledge base, making decisions and adding value (OECD, 2015_[1]).

A key purpose of creating value from data is to improve decision making and drive innovation. Data become valuable if used to improve social and economic processes, products, organisational methods, and markets. Data-driven innovation underpins many new business models that transform markets and sectors such as agriculture, transport and finance, driving productivity growth (OECD, 2015_[1]). More generally, data and data analytics are a key pillar of knowledge-based capital (KBC). KBC increasingly supports production in service and knowledge economies, and also includes intellectual property (e.g. patents, copyrights, designs and trademarks) and economic competencies (e.g. firm-specific human capital, networks of people and institutions, and organisational know-how) (OECD, 2013_[18]). Finally, value creation from data can be leveraged, for example by enhancing access to and sharing of data and thus fostering data reuse (see Chapter 2).

Identifying key challenges related to data

As data become a social and economic resource, including for value creation, decision-making, innovation and production, policy makers are facing a number of issues. Selected important issues include the value, ownership, flows and protection of personal data, as well as potential data concentration and divides.

It is difficult to assess the value of data in itself, given that value is essentially created when data are contextualised and analysed to derive information. In addition, the environment in which some data are used tends to be uncertain, complex and dynamic (e.g. research) (OECD, 2013_[17]). The value of data furthermore depends on their structure and the capacity to derive insights from them, notably analytic techniques and technology for data analysis, as well as prior knowledge and skills (OECD, 2015_[1]). Attempts to nevertheless impute the value of data remain imperfect proxies so far. For example, one estimate suggests that considerable consumer surplus is generated by digital products – and indirectly by the data used for and by them (Brynjolfsson, Eggers and Collis, 2018_[19]).

The concept of “data ownership” is controversial. The right to control access, copy, use and delete data – the main rights associated with the concept of data ownership – are affected in different ways, notably by different legal frameworks, e.g. copyright and related rights, and applicable sui generis database rights and trade secrets or, where personal data are concerned, privacy protection law (OECD, 2015_[1]; OECD, forthcoming_[14]). In practice, the intricate net of legal frameworks, combined with the involvement of multiple parties in the creation and reuse of data, including across national borders, motivates many stakeholders to rely on contract law as the primary legal means for defining proprietary rights related to data access and use (OECD, forthcoming_[14]).

Data increasingly underpin trade in the digital age and any measures affecting data flows are likely to have trade consequences, among others. Such measures may, for example, result from data-related regulation, such as local storage requirements, personal data protection agreements or trade agreements that cover cross-border data flows. A number of existing measures already make some cross-border data flows conditional or ban others altogether (Casalini and López González, 2019_[20]). Many of them concern personal data, in relation to which the OECD 1980 Recommendation of the Council concerning Guidelines Governing the Protection of Privacy and Transborder Flows of Personal Data recommends that: “Any restrictions to transborder data flows of personal data should be proportionate to the risks presented, taking into account the sensitivity of the data, and the purpose and context of the processing” (OECD, 2013_[21]).

Protecting data requires managing risk. The benefits of storing, using, accessing and sharing data come with potential risks that may arise from any of these activities, and risks need to be managed well to maximise benefits (OECD, 2015_[22]). This balancing act involves costs and legitimate private, national, and public interests, in particular the rights and interests of the stakeholders involved in producing and using data. Privacy and IPRs need to be protected and enforced; otherwise incentives to contribute data and to invest in data-driven innovation may be undermined, in addition to direct harm that may occur to rights holders (including data subjects) (OECD, forthcoming_[14]).

Data may also not be equally distributed. Concentration of data is visible, for example, in countries with many domestically hosted sites and high numbers of co-location data centres, often countries with a large population and uniform policies. Concentration is also present at sectoral and/or firm level, with some companies holding disproportionally more data than others. The same companies also tend to concentrate the capacity needed to create value – information and knowledge – from data. Information and knowledge asymmetries may in turn affect the distribution of power, with shifts: 1) away from individuals to organisations (including consumer to business, and citizen to government); 2) from traditional businesses to data-driven businesses; 3) from governments to data-driven businesses; and 4) from lagging economies to data-driven ones. These shifts in turn result in new divides, with implications for social cohesion and economic resilience (OECD, 2015_[1]).

National data strategies can help realise the potential of data, including through sharing and reuse. Strategies aimed to balance the issues mentioned above and achieve a social contract that unleashes the potential of data are uncommon today. However, some countries are in the process of developing such a strategy, and some data-related aspects are already addressed in open government data strategies as well as in national digital economy and/or security strategies, and others are on the verge of being addressed in emerging national privacy strategies (OECD, forthcoming_[23]). Building on these existing strategies, governments could consider developing consolidated broader data strategies as a comprehensive and coherent approach to leverage the potential of data for value creation while addressing the related challenges (OECD, 2018_[24]).

Key properties (“vectors”) of digital transformation and evolving business models

The use of digital technologies and data underpins digital transformation across many sectors and policy areas. To better understand the cross-cutting effects of the transformation, the OECD has identified seven “vectors of digital transformation” that describe key properties of digital transformation (OECD, 2019_[25]). The vectors of digital transformation offer an overarching perspective by describing the underlying and cross-cutting nature of the changes induced by digital transformation and their implications across different sectors and policy domains. As a result, they help overcome an often fragmented understanding of policy issues and facilitate a whole-of-government approach to shaping digital transformation (see Chapter 9).

Many properties of digital transformation, e.g. new sources value creation, impact business models and organisations. Firms that make use of digital technologies and data often compete in areas previously dominated by large incumbents. In some cases, new players create entirely new markets; in other cases, they shake up existing ones, driving structural change and incentivising traditional businesses to reinvent themselves. Recent business models that have emerged include those using online platforms and those combining online and offline features. The development of new payment mechanisms, which support digital transactions in a range of business models, represents another business innovation.

Scale, scope and speed

Digital technologies and data enable firms to create digital products or digitise existing ones, to digitalise business processes, to buy and sell online, and to implement new business and organisational models. These opportunities underpin the digital transformation of products, firms and markets. Three key properties of that transformation include economies of scale without mass, new economies of scope in digital environments, and speed and have a range of policy implications (Table 1.1).

1.1. Vectors of digital transformation: Scale, scope and speed
Vectors	Description	Examples of policy implications
Scale without mass	Core digital products and services, notably software and data, have marginal costs close to zero. Combined with the global reach of the Internet, these products and the firms and platforms that use them can scale very quickly, often with few employees, tangible assets and/or no geographic footprint.	The scale effect of being digital may allow the rapid acquisition of market share – which may be fleeting – suggesting that policies should ensure that barriers to entry and innovation are low, and adjust size-based approaches such as de minimis thresholds and categorisation based on number of employees.
Panoramic scope	Digitisation facilitates the creation of complex products that combine many functions and features (e.g. the smartphone) and facilitate extensive versioning, recombination and tailoring of services. Interoperability standards enable the realisation of economies of scope across products, firms, and industries.	Policies may need to span multiple policy domains, requiring co-ordination across historically separate issue areas and a more multidisciplinary perspective. This may argue for high-level principles as opposed to narrow rules, a shift from strict harmonisation to interoperability, and the convergence of policy oversight authority.
Speed: Dynamics of time	Digitally accelerated activities may outpace deliberative institutional processes, set procedures and behaviours, and limit human attention. Technology also allows the present to be easily recorded and the past to be probed, indexed, repurposed, resold and remembered.	Guiding policy principles may be preferred to specific rules that may be quickly rendered obsolete. New approaches such as the use of regulatory sandboxes and the exploitation of data flows and big data analytics may both accelerate and enable more iterative and agile policy making.
Source: OECD (2019_[25]), “Vectors of Digital Transformation”, https://dx.doi.org/10.1787/5ade2bba-en.

A first characteristic of many firms that sell digital products is the ability to quickly reach large scale without accumulating much mass. Unlike physical products, which tend to have high fixed costs and substantial marginal costs that decline with scale, digital products tend to have mainly fixed costs and low, close to zero, marginal costs. This characteristic, combined with the global distribution enabled by the Internet, allows successful firms and platforms to scale quickly, internationally and sometimes with very few employees or tangible assets, and thus “without mass” (Brynjolfsson and McAfee, 2014_[26]). While no firm can scale entirely without mass, digital products allow firms to go global without establishing many (if any) plants or hiring many employees. This is in stark contrast to brick-and-mortar industries, where global expansion requires at least some physical presence.

A second characteristic are new economies of scope in digital environments. Once viewed as a benefit realised by conglomerates that could support many product lines by sharing common costs such as legal, finance, accounting, and marketing, or through vertical integration, economies of scope in the digital era come with the ability to categorise, code and store information in standardised digital form, which provides the basis for efficient interaction and reduces transaction costs (Goldfarb and Tucker, 2017_[27]). In turn, firms can tailor digital products to individuals in near real-time, establish and maintain customer relations over time, and sell different products, while blurring sectoral boundaries (e.g. a firm operating in retail, ICT services and fulfilment/logistics).

Economies of scope also reflect the capacity of digital technologies to combine many functionalities through efficient combination, integration, miniaturisation and virtualisation. This in turn facilitates combinatorial innovation and engineering which allows functional expansion, such as in a smartphone that typically combines telephony, navigation, photography and music, and allows people to add a host of other applications all in one device (Varian, 2017_[28]). As data-driven business models proliferate across sectors from agriculture to finance to transportation to retail, data savvy firms have a comparative advantage, enabling and inducing them to broaden their scope and expand to additional sectors either as new entrants or through acquisitions of existing firms.

A third characteristic is speed, expressed by accelerated economic and social activity: markets clear faster, ideas spread more quickly, the time buffer associated with distance shrinks, as does the time it takes to engage and develop a community or to bring a product to market. Advantage increasingly goes to first movers and fast followers, and to agility supported by rapid, iterative learning. This underpins three business practices that have been associated with the digital era: 1) those that promise to “move fast and break things” (Taplin, 2017_[29]); 2) those that achieve scale before profits,4 facilitated by the near zero marginal cost of digital communication and information sharing; and 3) those that launch an idea before it is perfected with the assumption that iterative learning will come from its use in the market. These characteristics motivate firms to learn quickly – including how best to exploit a slow-moving policy environment.

1.4. Business models based on digital payment innovations

While many firms benefit from either scale, scope or speed, some benefit from any combination of the three. For example, the financial sector was an early adaptor of digital technologies and has seen significant innovation in data-driven business models recently. Digital solutions have started “unbundling” many of the functions previously carried out primarily by banks, such as payment – notably via cash, debit and credit cards or wire transfer – credit, trading and securitisation (OECD, 2018_[30]; OECD, 2018_[31]).

Digital payment innovations, like mobile money (e.g. M-Pesa), mediated by mobile network operators, are becoming more established, particularly in less-developed financial markets. Some cryptocurrencies promise transparency and immutability of transactions, and digital wallets increasingly enable connected devices to withdraw and transfer money on demand, thereby enabling frictionless payments in the real world.

Such solutions can be scaled up without investing in credit card readers, ATMs or physical banks, including across borders (scale); they can be embedded in online point of sales as well as increasingly in physical stores (scope); and some can be deployed fast and widely, often delivering faster service than traditional payment solutions (speed).

Ownership, assets and economic value

The advantages of scale, scope and speed enabled by digitisation and digitalisation of products, processes and organisations create incentives for firms to invest in intangible assets and new sources of value. Such firms can be digital pure players that start and run their business entirely online. Traditional firms also increasingly invest in intangible assets to enhance their physical products with digital features and/or ancillary services. Finally, some firms that started entirely digital are now expanding into the physical world as well. As a result, firms are increasingly exploiting intangible capital and tapping into new sources of value creation, with implications for policy (Table 1.2).

1.2. Vectors of digital transformation: Ownership, assets and economic value
Vector	Description	Examples of policy implications
Intangible capital and the new sources of value creation	Investment in intangible forms of capital like software and data is growing. Sensors that generate data allow machinery and equipment (e.g. jet engines, tractors) to be packaged with new services. Platforms enable firms and individuals to monetise or share their physical capital easily, changing the nature of ownership (e.g. from a good to a service).	Policy makers may want investment incentives to be more aligned with the economics of digital innovation and production (e.g. R&D, data, intellectual property). The ability to efficiently market services derived from capital equipment (as opposed to direct investments) may have implications for incentives to invest as well as measures of investment and productivity.
Note: R&D = research and development.
Source: OECD (2019_[25]), “Vectors of digital transformation”, https://dx.doi.org/10.1787/5ade2bba-en.

Since the mid-2000s, a growing share of business investment consists of intangible assets rather than traditional physical capital (OECD, 2013_[32]). Investments in intangible assets have grown quickly and now match or exceed traditional capital in a number of developed economies (Corrado, Hulten and Sichel, 2006_[33]). As they are intangible, assets in the form of know-how or business processes can be wholly or partially digitised and – encoded in data and software – enable firms to adopt new forms of organisation, new sources and processes of value creation, and new business models.

Investment in intangible assets and digital products have long started to deliver returns. While only a decade ago many viewed such investments as a long shot, firms selling digital products have in more recent years become the most valuable companies on the globe. In 2018, seven of the ten largest companies in the world gained much, if not all of their revenue from digital products, and six out of the ten most valuable Internet firms were digital pure players that either operated an online platform, sold software or provided digital financial services (Meeker, 2018_[34]).

More traditional firms selling physical goods, as well as capital owners, can also tap into new sources of value creation. For example, firms such as Rolls Royce and John Deere use sensors embedded in their tangible capital goods (e.g. jet engines, tractors) to collect and use the data about the performance of the equipment and the conditions of its operation; this enables them to provide ancillary services, often sold in a package with the good (OECD, 2017_[6]). Furthermore, owners of assets like real estate, cars and computing power can increasingly utilise their capital by providing access to their assets and monetising it as a service via online platforms.

Another example is digitised factory floors and production process that incorporate a “digital twin” that operates in parallel to the physical process (OECD, 2017_[35]). This enables the collection and analysis of data that improves the performance of the production process. Plant operators can optimise the control of the plant to increase efficiency, make informed decisions regarding trade-offs between performance and durability, assign loads and line-ups, perform the maintenance tasks at the right moment, head off costly problems before they occur, and explore the future through simulations.

1.5. Business models that combine online and offline features

Many traditional firms are increasingly moving online and combining both digital and physical components. While traditional firms are going digital, some firms that started online are now moving in the other direction. This extends beyond traditional firms simply having a website; instead it relates to viewing online environments as a seamless extension of the brick-and-mortar store and vice versa. On the one hand, traditional retailers make use of websites, mobile applications, self-check-outs, electronic kiosks and smart shelf technology; on the other hand, online retailers are starting to build digitally enhanced physical stores, removing frictions from traditional purchase processes and offering the option to “click and collect”.

In turn, consumer behaviour is changing. For example, consumers may research a product online before purchasing it in brick-and-mortar stores, while still reading reviews and comparing prices online. Similarly, other firms blend online and offline elements to sell goods of variable quality (e.g. fruits and vegetables) or goods that require a specific fit that is otherwise difficult to judge online (e.g. bespoke clothing) (OECD, forthcoming_[36]).

Relationships, markets and ecosystems

Digitisation and digitalisation would not be game-changers without the Internet. The Internet allows digital interaction, relationships and movement of value to take place at any distance and time; it enables markets to migrate or to be created from scratch online; and it facilitates the creation of ecosystems featuring multitudes of often interdependent actors, communities, products and markets. The transformation of space, empowerment of the edges, and platforms and ecosystems all have implications for policy (Table 1.3).

1.3. Vectors of digital transformation: Relationships, markets and ecosystems
Vectors	Description	Examples of policy implications
Transformation of space	Thanks to their intangible and machine-encoded nature, software, data, and computing resources can be stored or exploited anywhere, decoupling value from borders, and challenging traditional principles of territoriality, geographically based communities and sovereignty. This separation creates opportunities for jurisdictional arbitrage.	Policies relying on geographical specifications like nexus, rules of origin or defined markets may need to be revised, to consider other points along the process of value creation and distribution (e.g. location of value creations vs. value delivery). This separation of value creation from use increases the need for policy interoperability between countries and regions.
Empowerment of the edges	The “end-to-end” principle of the Internet has moved the intelligence of the network from the centre to the periphery. Armed with computers and smartphones, users can innovate, design and construct their own networks and communities through mailing lists, hyperlinks and social networks.	Public policies need to consider reorientation away from central (large institutions) toward more granular units like individuals. This includes policies ranging from digital security to labour and social policies.
Platforms and ecosystems	Lower transaction costs of digital interactions reflect the development not only of direct relationships but also digitally empowered multi-sided platforms, which in turn contribute to further reducing transaction costs in many markets. Several of the largest platforms essentially serve as proprietary ecosystems with varying degrees of integration, interoperability, data sharing and openness.	Public policies need to consider the market dynamics of online platforms, which may increase efficiencies but also re-intermediate and concentrate activities, which may have implications for maintaining sufficient competition. Governments may also need to rethink the provision of public services to take advantage of platforms.
Source: OECD (2019_[25]), “Vectors of digital transformation”, https://dx.doi.org/10.1787/5ade2bba-en.

The Internet affects previously existing networks, triggers a migration of intelligence from the centre to the edges, and drives convergence. Thirty years ago networks were specific to the type of service or content they provided. For example, switched telephony networks were used to transmit voice while broadcasting networks were used to transmit video. Such networks had an intelligent centre but “dumb” end-user devices like an analogue phone or a TV. The Internet changed this through the “end-to-end” principle that is at the heart of the Internet protocol5 . Intelligence of the network has moved from the centre to the edges where “application-specific functions reside in the end hosts of a network (e.g. a smartphone) rather than in intermediary nodes” (Saltzer, Reed and Clark, 1984_[37]; Estrin, 12 August 2015_[38]). The Internet also allows transmitting different kinds of data and information, for example text, voice and video, driving the convergence of previously distinct networks.

As the Internet becomes more pervasive and the cost of its use declines, individual users can communicate with many others, in effect setting up new networks built on the Internet. Such “many-to-many” communications sidestep other hierarchical or “command-and-control” structures of processing information. Just as the industrial revolution led to the invention of the modern limited liability corporation, so might the digital era lead to new, flexible forms of organisation, configured from an array of quasi-independent small enterprises and individuals. Decomposing and recombining smaller components of value may further reduce the distinction between economic categories such as business and consumer, work and leisure, and home and office.

This functional decentralisation leads to the empowerment and broadening of networks, markets, and communities and affects where power and influence reside as well as interactions among people, firms and governments. Reduced information asymmetries offer new opportunities for individuals and communities; regions can connect to global value chains; entrepreneurs can connect to potential clients, funders and suppliers around the world; and individuals can become publishers or journalists. But many-to-many communication and decentralisation also fragment control over information and erode the influence of traditional arbiters of information or “one-to-many” institutions such as newspapers, broadcast TV and radio, and governments.

While the shift of intelligence from the centre to the edges promotes decentralisation, online intermediation also creates opportunities for centralisation. In particular, online platforms provide intermediation on the Internet, enabling e-commerce, content distribution, search and storage services, and social networks (Box 1.6). On the one hand, online platforms promote decentralisation by lowering the barriers to participation, often furthering empowering the edges. For example, platforms like Amazon, MercadoLibre and Alibaba lower the cost of starting a business by providing fast, easy ways to set up online storefronts, reach customers and fulfil orders. On the other hand, platforms can also concentrate control in a proprietary service that owns the underlying technology, sets the ground rules for interaction, and collects data from and about users.

1.6. Business models using online platforms

While many definitions of an online platform exist, a consensus is emerging that online platforms are a “digital service that facilitates interactions between two or more distinct but interdependent sets of users (e.g. firms or individuals) who interact through the service via the Internet” (OECD, forthcoming_[39]). Online platforms are increasingly used to facilitate and structure online interactions and transactions, match supply and demand in markets for information, goods and services, and bring together one or multiple networks (also called “sides”) (OECD, 2016_[40]). For example, search engines help people find information, while also matching advertisers to users; ride-sharing platforms match passengers to drivers; social networks enable dialogue, content sharing, and commerce between individuals, businesses and advertisers; and e-commerce platforms match buyers and sellers.

In particular multi-sided platforms centralise online interactions, even if they happen independently within (seemingly) separate networks. Such platforms benefit from network effects: direct effects whereby the value of a service provided increases with the number of users, and indirect effects, whereby the number of users of one service increases the value of complementary services. By reducing information asymmetries and transaction costs online platforms can also make markets more efficient. This in turn allows firms that traditionally would “rather make than buy” (Coase, 1937_[41]) when information and input prices are uncertain to rather buy directly on the market.

Sources: OECD (forthcoming_[39]), Online Platforms: A practical approach to their economic and social impacts; OECD (2016_[40]), “New forms of work in the digital economy”, https://dx.doi.org/10.1787/5jlwnklt820x-en; Coase (1937_[41]), “The nature of the firm”, https://www.jstor.org/stable/2626876?seq=1#page_scan_tab_contents.

Beyond the emergence of online platforms, digital technologies are enabling the development of digital ecosystems and related business models. Such ecosystems include combinations of applications, operating systems, platforms, and/or hardware that interoperate in certain ways to enhance the user experience and/or aggregate data (e.g. Amazon’s Fire tablets, the Fire OS fork of the Android operating system, and interoperable apps and ebooks; Apple’s iPhones and iPads, their iOS operating system, and interoperable apps from the Apple App Store). Ecosystems can offer users ease of use, convenience, and a familiar look and feel with which they may grow comfortable, but they may also limit interoperability outside the ecosystem. While this can create advantages for businesses exploiting a model that stretches across a whole ecosystem, it may also raise users’ switching costs if and when a better product comes along, thereby helping incumbents to fend off market entrants and competition from one another.

References

Brynjolfsson, E., F. Eggers and A. Collis (2018), “Using massive online choice experiments to measure changes in well-being”, NBER Working Paper No. w24514, https://ssrn.com/abstract=3163281. [19]

Brynjolfsson, E. and A. McAfee (2014), The Second Machine Age: Work, Progress, and Prosperity in a Time of Brilliant Technologies, WW Norton & Company, New York. [26]

Casalini, F. and J. López González (2019), “Trade and Cross-Border Data Flows”, OECD Trade Policy Papers, No. 220, OECD Publishing, Paris, https://doi.org/10.1787/b2023a47-en. [20]

CISCO (2018), “Visual networking index: Forecast and trends”, https://www.cisco.com/c/en/us/solutions/collateral/service-provider/visual-networking-index-vni/white-paper-c11-741490.html (accessed 4 January 2019). [10]

CISCO (2017), The Byte Scale, CISCO, https://www.cisco.com/c/dam/assets/sol/sp/vni/qa_c67-482177-1.jpg (accessed 4 October 2018). [9]

Coase, R. (1937), “The nature of the firm”, Economica, New Series, Vol. 4, No. 16, pp. 386-405, https://www.jstor.org/stable/2626876?seq=1#page_scan_tab_contents. [41]

Corrado, C., C. Hulten and D. Sichel (2006), “Intangible capital and economic growth”, NBER Working Paper No. 11948, https://www.nber.org/papers/w11948. [33]

Drake, W., C. Vinton and W. Kleinwächter (eds.) (2016), Internet Fragmentation: An Overview, World Economic Forum, Davos, https://www.zora.uzh.ch/id/eprint/121102/1/WEF_FII_Internet_Fragmentation_An_Overview_2016.pdf. [42]

Estrin, J. (12 August 2015), Kodak’s first digital moment, The New York Times, https://lens.blogs.nytimes.com/2015/08/12/kodaks-first-digital-moment/?_r=0 (accessed 21 February 2019). [38]

Goldfarb, A. and C. Tucker (2017), “Digital economics”, NBER Working Paper No. w23684, https://www.nber.org/papers/w23684. [27]

Hofheinz, P. and D. Osimo (2017), Making Europe a Data Economy: A New Framework for Free Movement of Data in the Digital Age, The Lisbon Council, Brussels, https://lisboncouncil.net/publication/publication/143-a-new-framework-for-free-movement-of-data.html (accessed 21 February 2019). [12]

Kommerskollegium (2014), No Transfer, No Trade: The Importance of Cross-border Data, National Board of Trade, Stockholm, https://www.kommers.se/Documents/dokumentarkiv/publikationer/2014/No_Transfer_No_Trade_webb.pdf. [13]

Meeker, M. (2018), Internet Trends, https://www.kleinerperkins.com/files/INTERNET_TRENDS_REPORT_2018.pdf. [34]

Metodi, T., A. Faruque and F. Chong (2011), Quantum Computing for Computer Architects, Morgan & Claypool Publishers, San Rafael, CA. [7]

Moore, G. (1965), “Cramming more components onto integrated circuits”, Proceedings of the IEEE, Vol. 86, No. 1, pp. 82-85, http://www.cs.utexas.edu/~fussell/courses/cs352h/papers/moore.pdf. [2]

OECD (2019), “Vectors of digital transformation”, OECD Digital Economy Papers, No. 273, OECD Publishing, Paris, https://dx.doi.org/10.1787/5ade2bba-en. [25]

OECD (2018), Financial Markets, Insurance and Pensions: Digitalisation and Finance, OECD, Paris, https://www.oecd.org/finance/private-pensions/Financial-markets-insurance-pensions-digitalisation-and-finance.pdf. [30]

OECD (2018), Going Digital in a Multilateral World, OECD, Paris, www.oecd.org/mcm/documents/C-MIN-2018-6-EN.pdf. [24]

OECD (2018), OECD Equity Market Review: Asia 2018, OECD, Paris, http://www.oecd.org/daf/ca/OECD-Equity-Market-Review-Asia-2018.pdf. [31]

OECD (2017), OECD Science, Technology and Industry Scoreboard 2017: The Digital Transformation, OECD Publishing, Paris, https://dx.doi.org/10.1787/9789264268821-en. [35]

OECD (2017), The Next Production Revolution: Implications for Governments and Business, OECD Publishing, Paris, https://dx.doi.org/10.1787/9789264271036-en. [6]

OECD (2016), “New forms of work in the digital economy”, OECD Digital Economy Papers, No. 260, OECD Publishing, Paris, https://dx.doi.org/10.1787/5jlwnklt820x-en. [40]

OECD (2016), “The Internet of Things: Seizing the benefits and addressing the challenges”, OECD Digital Economy Papers, No. 252, OECD Publishing, Paris, https://dx.doi.org/10.1787/5jlwvzz8td0n-en. [3]

OECD (2015), Data-driven Innovation: Big Data for Growth and Well-being, OECD Publishing, Paris, https://dx.doi.org/10.1787/9789264229358-en. [1]

OECD (2015), “Recommendation of the Council on digital security risk management for economic and social prosperity”, in Digital Security Risk Management for Economic and Social Prosperity: OECD Recommendation and Companion Document, OECD Publishing, Paris, https://doi.org/10.1787/9789264245471-1-en. [22]

OECD (2014), “Cloud computing: The concept, impacts and the role of government policy”, OECD Digital Economy Papers, No. 240, OECD Publishing, Paris, https://dx.doi.org/10.1787/5jxzf4lcc7f5-en. [5]

OECD (2013), “Exploring data-driven innovation as a new source of growth: Mapping the policy issues raised by ‘big data’”, OECD Digital Economy Papers, No. 222, OECD Publishing, Paris, https://dx.doi.org/10.1787/5k47zw3fcp43-en. [32]

OECD (2013), “Exploring the economics of personal data: A survey of methodologies for measuring monetary value”, OECD Digital Economy Papers, No. 220, OECD Publishing, Paris, https://dx.doi.org/10.1787/5k486qtxldmq-en. [17]

OECD (2013), “Privacy Expert Group report on the review of the 1980 OECD Privacy Guidelines”, OECD Digital Economy Papers, No. 229, OECD Publishing, Paris, https://dx.doi.org/10.1787/5k3xz5zmj2mx-en. [21]

OECD (2013), Recommendation of the Council Concerning Guidelines Governing the Protection of Privacy and Transborder Flows of Personal Data, OECD, Paris, http://www.oecd.org/sti/ieconomy/2013-oecd-privacy-guidelines.pdf. [15]

OECD (2013), Supporting Investment in Knowledge Capital, Growth and Innovation, OECD Publishing, Paris, https://dx.doi.org/10.1787/9789264193307-en. [18]

OECD (2011), Quality Framework and Guidelines for OECD Statistical Activities, OECD Publishing, Paris, http://www.oecd.org/officialdocuments/publicdisplaydocumentpdf/?cote=std/qfs(2011)1&doclanguage=en. [16]

OECD (forthcoming), A Dynamic E-commerce Landscape: Developments, Trends and Business Models, OECD Publishing, Paris. [36]

OECD (forthcoming), Enhanced Access to Data: Reconciling Risks and Benefits of Data Sharing and Re-use, OECD Publishing, Paris. [14]

OECD (forthcoming), Online Platforms: A Practical Approach to their Economic and Social Impacts, OECD Publishing, Paris. [39]

OECD (forthcoming), “The road to 5G networks: Experience to date and future developments”, OECD Digital Economy Policy Papers, OECD Publishing, Paris. [4]

OECD (forthcoming), “Towards national privacy strategies”, OECD Digital Economy Policy Papers, OECD Publishing, Paris. [23]

Saltzer, J., D. Reed and D. Clark (1984), “End-to-end arguments in system design”, ACM Transactions on Computer Systems, Vol. 2, Issue 4, pp. 277-288, https://doi.org/10.1145/357401.357402. [37]

Siegler, M. (2010), Eric Schmidt: Every 2 Days, We Create As Much Information As We Did up to 2003, Tech Crunch, https://techcrunch.com/2010/08/04/schmidt-data/?guccounter=1 (accessed 4 October 2018). [8]

Taplin, J. (2017), Move Fast and Break Things: How Facebook, Google, and Amazon Cornered Culture and Undermined Democracy, Little, Brown and Company, Boston. [29]

United States Securities and Exchange Commission (1997), Amazon Annual Report 1997, https://www.sec.gov/Archives/edgar/data/1018724/0000891020-97-000603.txt. [43]

Varian, H. (2017), Measurement Challenges in High Tech, https://www2.census.gov/adrm/fesac/2017-06-09/Varian-Presentation.pdf. [28]

Weller, D. and B. Woodcock (2013), “Internet traffic exchange: Market developments and policy challenges”, OECD Digital Economy Policy Papers, No. 207, OECD Publishing, Paris, https://dx.doi.org/10.1787/5k918gpt130q-en. [11]

Notes

← 1. The industry standard setting process is led by 3GPP. A major milestone on the standardisation process of 5G was reached in June 2018, with the first phase of the standard intended for enhanced mobile broadband concluded. The second phase is expected to conclude in 2019 that will be designed to enhance the 5G Ecosystem for massive M2M and critical IoT applications.

← 2. One petabyte is roughly the amount of data produced by 3.4 years of continuous full high-definition video recording.

← 3. Capital goods include goods, other than material inputs and fuel, used in the production of other goods and/or services.

← 4. “Since inception, the Company [AMAZON.COM, INC.] has incurred significant losses, and as of March 31, 1997 had an accumulated deficit of USD 9.0 million. The Company believes that its success will depend in large part on its ability to (i) extend its brand position, (ii) provide its customers with outstanding value and a superior shopping experience, and (iii) achieve sufficient sales volume to realize economies of scale. Accordingly, the Company intends to invest heavily in marketing and promotion, site development and technology and operating infrastructure development. The Company also intends to offer attractive pricing programs, which will reduce its gross margins. Because the Company has relatively low product gross margins, achieving profitability given planned investment levels depends upon the Company’s ability to generate and sustain substantially increased revenue levels. As a result, the Company believes that it will incur substantial operating losses for the foreseeable future, and that the rate at which such losses will be incurred will increase significantly from current levels” (United States Securities and Exchange Commission, 1997_[43]).

← 5. “[…] every device on the Internet should be able to exchange data packets with any other device that was willing to receive them” (Drake, Vinton and Kleinwächter, 2016_[42]; Estrin, 12 August 2015_[38]).

End of the section – Back to iLibrary publication page