The two-sided digital economy

As we have seen above, the intermediary derives its effectiveness from its ability to treat the two sides of the market differently, of which it is the pivot. As soon as a service is launched and throughout the operation of this intermediation service, the operator makes every effort to develop the activity and eventually maximize its profit. One example is that of a payment card: consumers pay a relatively negligible fixed annual fee for such a card; traders, the other side of this market, pay a contribution that is proportional to the turnover paid to them by credit cards. The resources of the intermediary, who manages this means of payment, come from each of the two sides of this market (consumers and traders). This economic model is fundamentally different from the classic model, in which the trader’s activity benefits from economies of scale that gradually reduce the price of his collections as his turnover increases.

In many forms of intermediation, prices must be low, or even zero, in order to ensure the success of the intermediary, who may be able to increase the users’ contribution later when his position becomes comfortable. A balance must therefore be struck between increasing revenues and network deployment. The founder of the Amazon online sales platform, Jeff Bezos, first promoted the expansion of his network rather than the increase in revenue from sales; over the long term, his shareholders supported him: despite zero or even negative results, the platform finally established itself as an essential intermediary. In such a case, two concepts are contrasted at first glance: either investing to build long-term potential, or striving to make a profit in the short term! For most systems that offer a non-market service, such as search engines or social networks, the user knows full well that he does not pay for the service provided by the platform operator; but that the trace of the operations he/she performs is exploited by the operator according to rules that evolve over time, and which are not always the same for all operators.

Lastly, if the intermediary’s primary objective is to increase exchanges between the two sides of its clientele, it also enables them to establish relations within the same market component: on the same side, traders (or consumers) may try to cooperate with each other, for example. Or, on the contrary, they may be very competitive, opposing one another. For example, cooperation between players who play the same video game is frequent, especially in multiplayer games; similarly, the voluntary collaboration of developers of free software (so-called open source software) allows the constant improvement of such software, which is the product of a collaborative institution. On the other hand, competition occurs between companies that offer the same services to the same clientele or that seek to expand their clientele through the platform. These two modes of interaction (cooperation or competition between parties that are simultaneously present on the same side of the market) have an essential role for vehicle driver services: drivers, like passengers, are assessed by the other party (drivers are rated by customers and customers are rated by drivers); all compete with each other, both on the driver side and on the customer side of this dual market (see Uber).

Role of externalities

The competitive advantage or disadvantage of an intermediation platform is generally based on externalities¹⁰. These externalities can be positive or negative for third parties; they can be advantageous for some third parties and disadvantageous for others. Even more than in the classical economy, we can say that externalities play an essential role in the digital world. This is the case, of course, for Google, which is turning the economy of knowledge upside down in a way that is rather favorable to the majority of people (positive externality) but whose effect on the advertising market reflects a significant negative externality: the leader in search engines generated advertising revenue of nearly 60 billion dollars in 2014, to the detriment of other advertising media such as the written press or television!

However, data collected by intermediation operators are incredibly rich. They essentially reveal what is happening in the world through a continuous flow of information. They enable the holders of this information to maintain their knowledge of the field in real time, during the many interactions that pass through them. Thanks to these immense masses of data, they can extract new services that are sometimes independent of the initial service, precisely thanks to externalities. Non-monetized intermediation systems (a search engine, social networks, etc.) generate positive externalities that stimulate their activity. There is no reason why this should be different from monetized intermediation, such as transport (Triago) or accommodation (Airbnb), even if the intermediary’s income is based on a commission proportional to the amount of the transactions carried out.

Platforms like Uber now work closely with city administration. It is clear that such an operator will quickly be able to offer services similar to those provided by administrations – perhaps even more up-to-date? These services may be more effective because of their agility and the voluntary contribution of a large community of users, who thus become players in urban management. This new economy of resources, including but not limited to human resources, is also a strategic aspect of platforms. Externalities abound, especially for services that exploit geo-positioning: for mobile telephones, Android and iOS have far more sensors and much faster data processing than any territorial administration that would plan and organize travel or car traffic. On a completely different level, sales platforms can gather statistics on economic exchanges in real time, something that could replace or even contradict official statistics.

Abstraction in the cloud

Data collected by the platforms form a kind of mirror of the physical world; like any mirror, it is possible to view what is happening in the world without, however, having any interaction with it. Platforms are of course based on infrastructures, on data centers and on communication networks which are a significant investment¹¹, but without any connection with the intermediated world. It is therefore possible to establish and organize an interaction between two parties without any contact with them, or any link with the world they live in. Thus, Uber connects drivers and passengers in cities where this company has no other presence, except to arrange remote exchange between transport applicants and private drivers who are available to provide it. Uber has, however, a detailed representation of the city, of the vehicles that move in the urban space, of car traffic etc. This knowledge of the terrain allows Uber to match supply with demand at a faster pace than other carriers can.

Kept in a cloud (cloud computing), this abstract information about the territory is disruptive because it allows a service to be rendered without any visible presence on the ground¹². Worse still, intermediation in the cloud reaches a bewildering efficiency compared to the one their traditional competitors in the field are able to deliver¹³.

Law of concentration

Intermediation platforms did not exist until the 21^st century, so to speak; they have grown extremely rapidly; today they are disrupting many things. This is not surprising given the exceptional quality of their services. Their massive adoption in the world population gives them economic weight of the highest order: top world capitalizations are leaders in algorithmic intermediation: Alphabet, Alibaba, Apple, Google, Microsoft, Facebook, Tencent etc.

In almost all fields, one platform dominates its sector with a much higher number of users than its competitors¹⁴. This dominance of the larger platform is partly due to the fact that the power of the platforms follows “Metcalfe law”, which depends on the number of its users, in a quadratic manner¹⁵. It is also due to the fact that regional and local idiosyncrasies have a limited influence on this type of activity. Thus, Uber manages what could be considered as a universal transport service, which sparsely depends on local conditions and which is deployed over very varied territories, to such a point that it sometimes eliminates local operators that were installed long before it! This dominance of the leader is also due to venture capital, which invests heavily in a large number of projects, but only recovers its bet on the very few winners who are thus propelled to the top, eradicating a good part of the competition even if it existed well before their arrival. Moreover, concentration is not only capital-intensive; the number of users also follows a law of power: the result, for all economic sectors, is that the dominant platform is surrounded by a small number of operators, whose significance decreases to a modest size that allows many competitors to continue to compete. The concentration is also geographical, because few countries have been able to create large platforms; the United States are the first and the bay of San Francisco, the modern-world Venice, hosts most of them…

Two major transformations

As we have just pointed out, two transformations are rattling the world today: the digital revolution, on the one hand, and the climate revolution, on the other. Both of these disruptions are causing concern and profound change as a result of human activities. The digital revolution, which was made possible by new electronic technologies, seems to be heralding a new world. Conversely, the climate revolution, a negative externality of human activities and predominantly recognized since the first industrial revolution, seems to condemn certain practices. These two contemporary revolutions are probably less independent of each other than they seem at first sight; as digital technology is discovering new forms of organization and use of resources, particularly thanks to algorithmic intermediation. Beyond that, we can also see a feedback of resilience for our societies in digital technology [GRU 15].

The new digital economy primarily imposes a horizontal society, to the detriment of its previous vertical approach. This results from the continuous connection of each person to communication systems, in addition to the ability to interact with others through networks; for example, so-called social networks. The information systems preceding the Internet were very different: only point-to-point communication, especially by telephone, was available at the time, and only a tiny portion of the population was able to broadcast on mass media. Inevitably, communication that technology now allows is weakening vertical media, in other words: the hierarchy of communication structures is changing, and the governance of our society is being affected. Platforms play a decisive role in this change. Indeed, they contribute to destabilizing previous operators, following a rather simple process: the algorithmic intermediary operates a double shift of power, on the one side, downwards, which is favorable to the users of the services; and on the other, upwards, towards the platforms themselves, to the detriment of the established intermediaries. This double displacement:

• – gives full power to the users (consumer’s empowerment) through the services of greater abstraction, independent of the place where the service is rendered;
• – the upward displacement is facilitated by the platform’s ability to regulate intermediation between a very large number of players; it thus recovers valuable data and the associated business potential.

Historical intermediaries under threat

This double transfer of economic power is disturbing. An example that illustrates this phenomenon perfectly, which has already been cited several times, is that of taxi companies whose economic model is so damaged that their survival could be challenged. Beyond their economic model, the purpose of these many operators is collapsing. Another example, on a global scale: a significant proportion of the press has disappeared over the last decade¹⁶. Publishing houses are being challenged by self-publishing and digital distribution networks like Amazon, for example.

All players operating on two-sided markets are radically affected by one or more algorithmic intermediaries. Thus the real risk of being completely – or partially – overwhelmed because of digital technology. Why? Because the platforms offer personalized services of a higher quality than before. Let us return to books: a book can be self-produced by the author who has direct access to the various trades contributing to the manufacture and distribution of a book, which was formerly the prerogative of the single publishing house¹⁷. Production costs are reduced accordingly. This ability to connect directly with the professions that accomplish, produce and distribute a book illustrates the phenomenon mentioned above: it allows everyone to access the different professions in the book chain without an institutional intermediary. The institution which held this assembly ability and was thus the obligatory passage for the individual who needed it has been called into question. The professional monopoly collapses. Access to the essential resource, that of publishers’ expertise, has changed.

The strength of the platforms also lies in their detachment from the physical world in which they operate, a detachment that is strong for Google, weaker for Amazon storing and delivering objects everywhere in the world. The disruption capacity of both remains strong because their services are of a high level of abstraction, more abstract than those of traditional players who are disturbed by digital newcomers (e.g. publishers or booksellers facing Amazon); platforms are not constrained either by the safeguarding of what already exists, or by the regulatory or organizational constraints imposed on established players; the latter, however, will naturally try to preserve their position, at least in part! We understand the scope of this revolution for a number of sectors that are part of a two-sided market: producers, on the one hand, and consumers of services, on the other¹⁸. State institutions will undoubtedly not escape this threat, even if they have significant inertia as a result of their weak link with any market: universities, for example, would benefit from taking advantage of a development that could lead students to resort to education outside the institution in which they are enrolled! This phenomenon is already noticeable, since it already happens that young engineers are asked to post their curriculum on platforms (like GitHub, for example) beyond their academic career, stricto sensu.

Internet generalized to objects

Intermediation between humans will rapidly extend to non-human players, for example, transforming the energy environment: the single and reliable producer who supplies millions of consumers could evolve towards a distributed system that would rely on millions of less reliable producers, capable however of supplying millions of consumers; this would require that intermediaries put these productions in relation with consumers. Energy management, an essential resource, could thus be highlighted as an algorithmic intermediation. In order to confirm this hypothesis, one only has to look at the investments of the platforms which prepare the Internet of Things and home automation!

Integration is not easy to achieve, however. It is easy to exploit accessible resources intensively, but the quality and reliability of the previous vertical distribution must still be maintained. Ensuring this guarantee of supply is a decisive bet for the transformation of our societies. The replacement of vertical structures by horizontal structures requires thought into what the strength of vertical organizations used to be: in this case, it was the trust placed in the single producer! Without this trust, no society could function; but trust patterns sometimes change from one society to another; they can change over time [FUK 96]²¹.

One of the roles that states could play would be to create favorable conditions for this new confidence through a priori normative mechanisms and retrospective controls. However, trust could also result from cooperation among the multitude, for example through recommendation and assessment mechanisms²². It is quite remarkable that carpooling, an old utopian dream, has become an ordinary reality whose ramifications are spreading, given the obvious savings that it allows. What traditionally guaranteed the quality of a driver, the employer above him/her, is now guaranteed from below, in other words, by the users. The widespread integration of confidence-building mechanisms is a basis of the platform economy. The peer assessment or judgement mechanisms spreading to all sectors of society is not at random. They are witnessing a mainstream change!

Deep issues of concern

The platforms studied in this chapter have an operating mode that differs radically from that of traditional multinationals that penetrated territories from above; they were congruent with the existing local or national framework; in the event of conflict, they negotiated with the authorities, either visibly or in secret. In contrast, platforms enter territories from below; they directly affect citizens who adopt their services without seeking permission from anyone.

Platforms also challenge some of the rules that organize our societies, such as the protection of privacy, the intensity of state intelligence, the legitimacy of political decisions, institutional governance, and so on. All aspects that need to be considered in a long-term perspective, taking the depth of digital change and its irreversibility into account. The right to be digitally forgotten illustrates the complexity of the problem in an emblematic way²⁵. Although the validity of this right is not really called into question by current practices, its implementation puts forward many problems in digital society.

The decision of the Court of Justice of the European Union, forcing Google to deregister pages that contained personal information, at the request of interested parties in May 2014 and in accordance with EU management rules for personal data, raises countless practical problems and has given rise to heated controversy. This decision gives Google one more power: by letting a private operator judge the admissibility of an individual’s request, the European administration resigns because it would not have the means to do so itself! It also leads Google to be better informed of the requests of the interested parties, which amounts to enriching its knowledge on the users; it makes the search engine a little more arcane and also limits the right to inform, from a very European perspective! Google, who was of course very averse to this decision, nevertheless deployed the means to implement it. The dereferencing being only covered on European domains, the CNIL²⁶ asked for the effect of this decision to be extended to the whole domain (.com) at the international level in July 2015, which Google obviously refused. This case is at the core of the territorial jurisdiction issues that will dominate the debate between platforms and states in the coming years. While states have a historical legitimacy on their territory, platforms hold a power that conflicts with that of states. This is due to two things:

• – first, the direct link that these operators maintain with the populations and the quantity of information they hold on the territory (and on their population);
• – second, the fact that this direct link with the populations allows intermediation, which calls fundamental aspects of our social organization, such as labor law, into question.

Data is therefore a source of power. During the 19th and 20th centuries, the major Western powers had increasingly subtle methods of strengthening their knowledge of the population through censuses, whose issues became progressively more numerous and more precise, thanks to the progress of statistical techniques while economic development and quality of life also progressed ²⁷. Platforms now deliver solid information on the population and territories, information that seriously challenges official statistical institutes as well as data (fiscal, social, health or other) coming from administrations whose sources are slow to process and costly to exploit while platforms manage a continuous flow of information. At this point in time, it is clear that the relative share of platforms is increasing, to the detriment of that of public institutions.

The geographical distribution of the platforms is uneven. The United States has most of the global operators that hold a significant share of this activity in many countries. China has close to a third of the world’s major platforms, but its area of influence remains regional to this day. Some small digital powers also have a strong national activity. Europe, for its part, has no major platform; it basically depends on the major American operators mentioned several times in the preceding pages. For European countries, the tensions induced by the platforms therefore raise conflicts with foreign companies (notably outside the European Union). And although some European issues are similar to those found in the United States, for example on labor law or the protection of privacy, others emerge from a profoundly different perspective of the world: Europeans sometimes invoke their own values. However, the absence of proper local intermediation might eventually be added to Europe’s harsh failures!

By the number of their users, by the speed of their adoption by our populations and by the undeniable success of their services, intermediation platforms possess actual legitimacy in many territories. The prohibitions and lawsuits that are brought against them do not hinder their growth; these incidents leave them almost unharmed. The same will undoubtedly apply to the decision to ban the Uber Pop service in France, for example, because it is very difficult for the legislator to define a framework that will allow these exchanges to develop in compliance with tax and insurance rules that are poorly adapted to the new economy. Moreover, platforms that have a very strong capacity for influence are able to direct the result of an election on any territory²⁸. In addition, they are actively involved in elections for new uses, as was the case in San Francisco in November 2015, over a proposed law on short-term leases that Airbnb helped to defeat²⁹. This is a reality of democratic life of which the consequences are, at this stage, difficult to anticipate.