Digital Platforms vs Utility Infrastructures
Regulatory implications
Traditional utility infrastructures such as energy water and train transportation were once developed and operated by single companies – often with the support of the state. Why was that the case? Because these utility infrastructures necessitated high upfront investments and fixed costs, access to specific geographical spaces, and so on all of which were barriers to entry for other companies to enter the market. As such, these companies became natural monopolies. It was not beneficial to any government to change the state of the market because that would mean incurring additional costs to serve their population with basic utilities like energy and water. If we look at the energy grid for example – there are a lot of fixed costs involved to build and maintain the grid network. If another company was to set that up, it would cost billions. It is for that reason that it is more economically effective to have one supplier provide these basic utilities.
Of course, as history has shown, by the 1990s, it became clear that these natural monopolies became a state within a state – they set the prices and the terms of service and they in fact influenced government budgets. At that point in time, we had a lot of ex ante regulation to deregulate these natural monopolies and allow competition between multiple companies, bringing down prices and providing more fair services to end users.
Now some have argued that digital platforms are slowly becoming natural monopolies and should thus be regulated just like early utility infrastructures were. This is problematic because digital platforms are different from utility infrastructures. There are three characteristics that distinguish digital platforms from utility infrastructures, namely, they are built on a product agnostic architecture, they generate value both from the supply and demand side and they have very distinct governance rules.
Digital Platforms are Built on a Product Agnostic Architecture
First, an energy grid or a water infrastructure are built on very specific architectures and they are very much tied to the physical product they produce – that of energy and water. By extension, they have very specific product and service boundaries. In contrast, digital platforms are built on product agnostic architectures, most of which are cloud based, modular and layered. Now why is this important? Because they have no set boundaries. For example, Amazon Alexa, which is embedded in Amazon’s Echo device, it can offer voice activated streaming content from Amazon’s music library and Kindle audiobooks, as well as integrate music services from third parties such as Spotify. Amazon Alexa may also integrate with other third-party services such as smart thermostats, lighting switches and other smart home applications. It can also order food from Deliveroo, a ride from Uber and so on. Thus, the technological architecture of digital platforms enables generative business models with varied customer bases, across products and services, and with distinct revenue models and data aggregation strategies that can also be monetized. So, digital platforms are not contained within single industries or market sectors as in the case of utility infrastructures. Their market boundaries are permeable.
Digital Platforms Generate Value from the Demand & the Supply Side
The second point is about value creation which is connected to the first – it essentially refers to the business models of these digital platforms that are enabled by their underlying architecture. Unlike utility infrastructures that generate value for suppliers of services, building supply-side economies of scale, digital platforms rely on both supply and demand-side economies of scale. Without platform participants, including end users and third parties such as app developers and advertizers, the platform itself becomes less valuable. At the same time, however, these platform participants generate innovation across boundaries and thus are able to build new value creation and capture opportunities. Now what is interesting is that, it is exactly such open innovation that gives rise to new digital platforms.
We keep talking about the dominance of big tech companies but a very recent example of this is the gradual demise of Facebook and the generative spurt of new social media platforms such as TikTok. TikTok has exploded in the last year or with over 1.5 billion downloads and more than 500 million active users. So it's a massive platform. What's interesting is that they have now created a unit called TikTok for business in which they're engaging their creators and matching them with businesses for content around marketing and advertising and obviously they have a lot of insight into the size and the focus of those particular influencers that can be matched with different businesses.
By contrast, utilities infrastructures feature strong supply-side economies of scale, with suppliers capturing all the value for themselves. The products and services delivered through these infrastructures, such as electricity and water are standardised and homogeneous with no opportunities for differentiation other than cost. There are limited value creation opportunities for third parties relative to digital platforms, because utility infrastructure offerings are bound within a highly specific market. Innovation is mainly focused around the maintenance and improvement of existing physical infrastructures (e.g. upgrades to 5G telecom networks). In contrast, digital platforms benefit from constant innovation across boundaries and thus new value creation and capture opportunities.
Digital Platforms have very Distinct Governance Rules
The third point is that digital platforms have different governance rules and control mechanisms for orchestrating the production and consumption of services. These governance rules use both technical mechanisms such as API but also behavioral mechanisms such as curating content, access and use of platform resources. These governance rules are part and parcel of the technology architecture and market scope of the platform. The orchestrators of digital platforms need to protect their own interests in competition with other firms, while also allowing complementors who contribute to value creation on the platform to secure their interests. The way the platform firm balances these trade-offs is through enforcement of governance rules, which affect the extent of, for example, multihoming across platforms vs. exclusivity strategies; and convergence of market and competitive domains. These governance rules include gatekeeping through a set of boundary resources such as software development kits and standard interfaces. These governance rules also influence pricing strategies. Platforms use subsidies to deal with the chicken-or-egg dilemma to incentivise user and complementor participation, value creation and capture. They also bundle products through subscription, while also flexibly marking up star complementors (e.g. Amazon Prime Video subscriptions vs. premium content). Such pricing strategies depend on cross-side network effects and the respective demand elasticities for the different market sides. While on the surface, utility infrastructure suppliers use similar pricing strategies, utility pricing does not depend on cross-side network effects and demand-side elasticities, nor on the market power of complementors.
Digital Platform Regulation akin to Telco Prohibitions in the 1990s?
Although digital platforms such as the Apple App Store and the Amazon Marketplace have grown tremendously in the last decade or so, these are not ‘natural monopolies’. We need to differentiate between platform growth natural monopolies as conceived in the context of utility infrastructures. However, such practices as self-preferencing are definitely anticompetitive and they do raise issues of antitrust that resonate with those raised for early utility infrastructures. Let us draw some distinctions.
In the case of utility infrastructures, a firm owning an infrastructure asset, such as a telecom network, was challenged in relation to obligations it had to other
companies that needed to use its assets, and also compete with them. Natural monopolies in the telecom sector, for example, refused to supply those assets to new entrants, they implemented margin squeezes, they discriminated who to supply to and they tied products and services together. All of these were recognized categories of abuse of market dominance and they were called “discrimination”.
Now in the case of digital platforms this is called self-preferencing and it is observed when a platform favours its own products and services over the competing products of other entities. For example, Apple pre-installs native apps on all iOS devices and makes these the default option for users, by means of tightly integrating them with other iOS services. The set of API provided by Apple to developers always default to Apple’s pre-installed applications. For example, when an iOS user clicks on a link, the webpage opens in the Safari Browser, a song request opens in Apple Music, and an ebook opens with Apple Books.
Now at the surface, the two look very similar. The discrimination practices of early utility infrastructures are similar to the self-preferencing practices of digital platforms. The question we need to ask is whether self-preferencing is a standalone abuse. In other words, by treating such self-preferencing as abuse in and of itself, competition authorities seem to be making assumptions that self-preferencing kills the merits of competition – this however, may not necessarily be true, at least not in all cases. In other words, when Apple makes their API default to Apple-native apps, it is not necessarily disallowing users to choose other apps. And it has all the right to try to integrate Apple services across its platform to provide a better user experience to Apple users.
But this does not mean that it is acting unlawfully and the users have a choice. For example, despite that a song request by an app preferences Apple Music, users tend not to use Apple Music much – Apple Music only has 15% market share. Spotify is still the market leader with 31% market share. So, in this case, self-preferencing does not kill the merits of competition – a higher quality app with better user experience will be preferred by users. And for Apple users service integration still remains important.
However, in the case of Apple’s app tracking transparency policy, where users can choose to opt-out from third party apps tracking their data, BUT they cannot do so – they cannot opt out from Apple apps tracking their data, that is clear abuse. That is unlawful self-preferencing and clearly should be regulated. In this case, there is a clear impact on user choices as well as the business development of third party providers of services and products.
This is why we need both ex ante and ex post regulation. Ex-ante regulation could include a general prohibition against self-preferencing one’s products and services against those of third parties. But this needs to be complemented by ex-post regulation that could include specific provisions for third parties to complain against unequal or unfair treatment.
Access to Digital Platforms as Essential Facilities?
Some digital platforms like Google Search have become essential in our lives. But is ithis the same as the ‘essential facilities’ of electricity or water? Access to the Internet, including bandwidth and speed, is something that should be treated separately. There is a lively debate on Net Neutrality which will actually be impacted by 5G networks, as I discussed in another article.
Outside access to the Internet, however, it is dificult to convincingly argue that products and services offered on Facebook, Twitter, TikTok or even the App Store and Google Play can be considered essential facilities. A facility is essential if no reasonable alternatives are available and duplication of the facility is not feasible due to legal, economic or technical obstacles. So access to bridges or ports or other physical infrastructures are essential facilities because there are no other alternatives in their respective localities. But this is not true for Facebook or TikTok, for example. There are alternatives to those and there are no legal economic or technical obstacles for accessing those. So in the case of end users – i.e. non-business users, there are no issues here.
In the case of business users, however, we come back to ideas around discrimination and self preferencing. So, with Apple’s App Tracking Transparency, other businesses including data aggragators, advertizers, and app developers are not treated fairly. In those cases, yes, we have anti-competitive practices, because Apple collects data from users, whereas others cannot. Users can opt out from the use of their data by third parties. A fair solution would be to allow users to opt out of all data collection from all apps, whether Apple native apps or those from third parties.
So, if we want to argue for fairness and equal access to all we should provide end users with a choice to either opt in or opt out across all apps and services. It should be explained to end users, however, that by opting out, many of the services that are customized to their needs would also disappear. So, there are trade-offs involved for the end user. But these trade-offs should impact all business users equally if we want to achieve fairness.