Platform research in focus: decentralisation and network effects

“As they say, if you want to know where you’re going, you have to know where you’re coming from”, says Joost Rietveld. Considering the many new avenues of innovation – and the speed with which platforms and technology are moving – the adage is well chosen. As Associate Professor at the UCL School of Management, Joost actively explores questions of interorganisational dynamics, digitisation and technology strategy while keeping abreast of the wider platform research landscape. “Research on platforms really is booming”, he says. “There have been over 600 papers over the last 30 years or so. But 332 of those were published in the last five years alone, and 2023 is on track to be the best year yet in terms of academic studies on platform competition.”

Major topics in platform research

Today, platform research is conducted by scholars across a range of disciplines, such as management and organisation, information systems, economics and marketing. This is causing platform research to grow not only in breadth, but also in depth, says Joost. “There’s a lot of work being done on network externalities and network effects, but we can dig a little bit deeper and look into other topics as well,” he tells us. To better understand the nature of current research, he recently conducted a literature review and analysed which topics are drawing the most attention.

There are four major focal areas of platform research that Joost has identified: network effects, corporate scope, orchestration and heterogeneity. The first deals with a fundamental characteristic of the platform business model and how to leverage it effectively. Joost shares some of the emerging research questions around network effects: “How do you scale up a platform? What is the difference between indirect and direct network effects? Some platforms only attract end users, whereas others connect businesses to end users – so how do you get those indirect network effects going?

The topic of corporate scope looks at the range of business activities or features that a platform might adopt; for example, Apple integrated flashlight functionality directly into the iPhone once it became obvious that flashlight apps were proliferating on its App Store. “Corporate scope could also be Google getting into healthcare or Microsoft buying a video game publisher to expand its activities”, Joost explains.

“Orchestration is all about how you manage your ecosystem”, he continues. This kind of research considers aspects of platform governance that aren’t related to pricing – in other words, how to coordinate users on different sides of the platform using non-economic means.

Heterogeneity as a research topic asks, according to Joost: “How do platform users differ from each other? How do platforms differ from each other and what are the implications of that for strategy management and market penetration?”

 Video games: the perfect research setting

While all of these topics are rooted in economics research on network effects and pricing, interest is growing in platform-specific strategy, which leads researchers to look more closely at specific contexts. Scholars are studying a range of different settings, but one has emerged as the most popular by far: video games. “Video game consoles are very often referred to as the canonical two-sided market setting, because you have gamers on one side [and] game developers on the other side”, Joost explains. “Game consoles are being sold at a loss to consumers in order to ramp up that user base, and hopefully that attracts game developers. There are network effects; markets are somewhat tippy; there are generational breaks. It’s a really good setting to study all sorts of platform dynamics.”

At a close second are mobile app stores, which have become another setting of interest for scholars, Joost adds: “If you’re coming across academic research on two-sided markets, it’s very likely to either have been done on game consoles or mobile app stores.” Still, there is a growing number of studies looking at a variety of industries and settings, such as ride-sharing services, Airbnb, Amazon and other e-commerce marketplaces, crowdfunding platforms and pre-digital platforms like newspapers. “I’ve also come across papers about beauty salons, Chinese novel writing platforms and even dog-sitting platforms”, Joost says. “If you have an interest in academic research on platforms, there’s probably at least a couple of papers out there that look at your particular industry.”

Fostering public interest in platform research and increasing access to it are among Joost’s other projects, and he’s created two different resources that are publicly available. The first is platformpapers.com, a searchable reference database of platform research papers; the other is the Platform Papers blog, published on Substack. Every month, Joost invites a guest scholar to share and contextualise their research on the Platform Papers Substack. As the body of platform research continues to grow, Joost hopes that the discipline’s findings will reach a broader audience. “I encourage [everyone] to start consuming this research to deepen your intellectual base when it comes to platforms and platform competition”, he says.

A deep dive on decentralisation

Growing research may be the defining takeaway of where we’ve been, but the question of where we’re going remains. Jean-Philippe Vergne, Assistant Professor at the UCL School of Management, researches industry evolution and organisational adaptation, and he’s interested in a topic that is firmly at the leading edge of platform design: decentralisation.
“Today we’re very familiar with what I would call corporate platforms”, Jean-Philippe says. “They’re essentially platforms on top of which a corporation sits; for instance, Meta is one such platform corporation […that] owns, designs, controls and operates one or more digital platforms such as Facebook, Instagram, WhatsApp, et cetera.” In Jean-Philippe’s view, corporate platforms like Meta are also examples of centralised platforms. “There’s always going to be a CEO sitting on top of the platform corporation who’s in a position to reverse any decision that’s made by anyone else in the platform ecosystem”, he notes.
“By contrast, what I would call decentralised platforms […] would be a setup where there is no such platform corporation sitting on top”, Jean-Philippe continues. “Imagine a digital platform that can run, operate, grow [and] get millions of users without having a CEO, shareholders, headquarters, its own equipment and computers or even any employees. Can we make that happen?” From his perspective, the existence of Bitcoin as a platform for transacting and generating value already proves that the answer is yes. He points out that Bitcoin doesn’t have any of the operating components of a typical platform, such as a CEO or employees – but it does have dozens of users and market capitalisation of around half a trillion dollars.
With Bitcoin providing a clear example of a different type of platform model, the opportunity arises to define a new taxonomy of platforms.“We’ve got these two very distinguishable sets of entities: corporate platforms that are centralised, […] and then, on the other hand, a very different kind of beast [in] decentralised platforms”, Jean-Philippe says, also noting that either type of platform can differ in its degree of centralisation or decentralisation, respectively.

Centralised or decentralised platforms – and why it matters for regulation

It’s easy to underestimate the importance of this distinction between centralised and decentralised platforms, but the real-world regulatory implications can’t be overstated. “The big issue here is to understand how important this matter is to society and to regulators across jurisdictions”, says Jean-Philippe. “We have examples in the United States where regulators [have] said that if a digital platform is issuing digital assets, but the platform itself is decentralised enough, then maybe those assets shouldn’t be considered securities [and] should be exempt from regulations on securities.”
Decentralisation can also impact a platform’s liability and responsibility for governance, he continues. “[It’s a factor in] determining whether a platform should be considered a publisher or an intermediary that’s just enabling decentralised content sharing. That can have a big impact […] for regulations involving free speech.” And thirdly, decentralisation can influence a platform’s pathway to growth. “[It] matters whether a platform can be decentralised enough so as not to be considered a business entity that requires incorporating. That can have profound implications for the legality of the business models that are implemented”, Jean-Philippe tells us.

 Platform boundaries and the dispersion of authority

The stakes for decentralised platforms may be clear, but the criteria for what defines decentralisation is still up for debate. Jean-Philippe thinks that it comes down to an analysis of who the platform’s actors are and how much influence they have. “Defining the boundaries of your digital platform is fundamental for understanding how decentralised it is or not”, he says. “Platform boundaries should be drawn around the contributors who have authority over the operations of the platform.”
Jean-Philippe returns to Bitcoin as an example. In this case, the contributors in question are the users who have authority over Bitcoin. Wallet users control the key to their own wallet and are able to move funds, miners create blocks and order transactions and software developers update the code that runs the Bitcoin infrastructure. All three groups of contributors have certain types of authority over how the platform operates and are therefore all within the platform’s boundaries, Jean-Philippe says.
“What most people mean by decentralisation is the dispersion of authority within the platform”, he continues. “[Firstly, having authority] means that you have access to information that’s relevant to operating in the platform. And second, you are able to contribute to the important decisions about the operations of that platform.” To illustrate, Jean-Philippe lays out a graph with two axes, resulting in a grid with four different possible scenarios. The horizontal axis represents information dispersion, while the vertical axis indicates decision-making dispersion. These two factors combine to confer or restrict authority.

Decentralised platforms like Bitcoin are therefore located in the bottom-right quadrant, which represents maximum dispersion of authority on both axes among all users of a platform. However, even within a decentralised platform, different types of decisions demand certain decision-making rights and come with varying levels of responsibility. Jean-Philippe illustrates further: “On Bitcoin, [there are] about 15 million agents who can decide to sign a transaction. They’re called wallet users. There’s about 15,000 agents who can decide to accept or reject a new block of transactions. They’re called full nodes. And there’s about 30 agents who can decide to order pending transactions inside those blocks. They’re called mining pools.” All of these contributors have some decision-making power, but authority isn’t equally distributed throughout the Bitcoin platform. “You can see just how dispersion of decision-making can vary within a so-called decentralised platform such as Bitcoin”, Jean-Philippe notes.

This provides more context for how Jean-Philippe goes about analysing whether a platform is categorised as centralised or decentralised. “Once you understand the minimum and the maximum [amount of dispersion]”, he says, “you can really quantify just how much decentralisation you have in your digital platform.” Using a mathematical formula to calculate dispersion across a group of users, Jean-Philippe represents minimum dispersion as 0% and maximum dispersion as 100%. He explains: “[When it comes to decision dispersion,] 100% would mean that every agent who’s part of the Bitcoin platform can contribute to all the decisions. 0% means only one agent can do that.” By mapping this metric over time, researchers – or other platform leaders – can track the development of dispersion on a platform. “In 2009, [Bitcoin started out] very high in decision dispersion, around 80%, and then there was a big drop. In recent years, it started increasing again”, Jean-Philippe tells us. The same exercise can be done for information dispersion; the combination of these two aspects indicates how a platform is distributing authority among its users.

Applicable, valuable insights for platform strategy

The purpose of this kind of analysis isn’t simply to define decentralisation or inform policymakers. It’s also highly valuable to platform practitioners and entrepreneurs, according to Jean-Philippe: “Once you have this data, you can look at the impact of authority dispersion on interesting metrics: how fast you can grow as a platform, how fast you can attract complimenters, how fast you can increase your market valuation, how [authority dispersion] affects various types of risks.” In Jean-Philippe’s Bitcoin example, he is able to map authority dispersion against price changes or the number of cyberattacks on the platform. He can also use these metrics to compare Bitcoin with other similar cryptocurrencies or with more centralised platforms, such as the SWIFT payment network.
Digging into data can be fascinating on its own merits, but for Jean-Philippe, the goal of collecting and interpreting this information ultimately comes back to its real-world application. “Platform decentralisation can and should be measured, and the data on this decentralisation is a must-have for policymakers, a should-have for platform developers and executives and a nice-to-have for consumers who want to assess their risk when they deal with particular businesses”, he says.

The pros and cons of decentralisation

As we learn more about decentralisation, it naturally raises questions about the pros and cons of this type of platform model. In Jean-Phillipe’s assessment, decentralisation has both. “A big advantage of platforms that are decentralised is that they can make a credible claim of platform neutrality”, he says. This means that the platform can attract contributors – or complimentors that are supplementary to the platform ecosystem – and credibly prove that those contributors won’t have the rug pulled out from under them in the future, as Jean-Philippe puts it. Contributors want to be confident that the rules of the game won’t be changed unilaterally, he says, especially if they’re relying on the platform to earn their living. Decentralised platforms, by nature, limit the scope of change to their own infrastructure unless there’s broad consensus among their contributors. “I think there’s a number of markets and industries where this can create tremendous value, but not always and not everywhere”, says Jean-Philippe. 

“On the other side, I think centralised platforms [also] have a strong advantage”, he adds. “What we are seeing in the data is that they’re just faster at innovating. When you have a high level of authority dispersion, it can take a while to be able to make a decision, because you’re engaging in these very lengthy consultative processes. You want input, you want everyone to have a say, but sometimes the market is moving really fast and you want to be able to make a quick decision. And so that can slow you down.”

Moving slowly may not always be a negative, however. “In the case of Bitcoin, the fact that it’s very slow can be an asset. Given what Bitcoin is trying to do, I think the lack of innovation – the inertia – is a good thing. But in other markets, it’s bad – it’s hurting platforms that are too decentralised”, Jean-Philippe thinks.

Healthcare platforms: a case study in the power of network effects

In contrast to Jean-Philippe’s focus on management and organisation, Xu Zhang specialises in developing effective marketing strategies for digital platform markets. “Overall, my research is about helping digital platforms to design their pricing structure, as well as [to decide] what kind of information that you want to collect from consumers and what kind of information that you want to share with future consumers [to] encourage customer engagement with the platform and, at the same time, improve company revenue”, she explains.

An Assistant Professor of Marketing at the London Business School, Xu has been working on a large project about the evolution of digital healthcare platforms, which she uses as a case study to demonstrate lessons and insights that can benefit platforms across many different sectors. “We’ve been seeing digital healthcare platforms emerging all around the world in developed countries as well as in developing countries,” she tells us. “[We’re] spending a lot of money on healthcare, ranging between 5% to 20% of the GDP, and with the current ageing population, that number is likely to grow even higher”, Xu continues. “But despite that high spend, […] a significant part of the world’s population still lacks essential access to healthcare services.” She adds that the WHO has proposed using digital platforms as a solution for expanding global access to healthcare, thanks to a unique feature of the platform business model. “Compare digital platforms with the traditional business model in the healthcare sector – [the key thing with platforms] is the existence of network effects”, she says.

A refresher on network effects

Network effects may be a familiar term to platform experts, but it is often a new concept for fresh entrants into the world of platforms, particularly when they’re coming from an established sector like healthcare. “When we talk about online healthcare platforms, [network effects] can actually apply to very different parties in this ecosystem”, Xu summarises. “The healthcare provider side may include hospitals, doctors and health insurance companies. On the healthcare recipient side, it may include patients, consumers and insurance clients. The online platforms who sit in the middle as intermediaries facilitate this matching process, the information-seeking process and the transaction process between the different parties.” Network effects are clearly observable in this context, she adds: “The value of the platform to a hospital depends on how many patients have already joined.”

In healthcare, as in most other sectors, there are two major focal points that emerge among the various KPIs that platforms may be measuring. “Many of those metrics are related to profit and another set of metrics is related to growth”, says Xu, noting that platforms need to strike the right balance in how they prioritise these metrics. “In the long term, those [two areas are combined into a] single objective for a successful platform. However, in the short term or the medium term, those two objectives are very different”, she says. “If you want to prioritise profit, then likely, you can’t subsidise participants that much – you can’t give them a lot of incentives, so you are going to suffer on the growth front.” The same can be true in the reverse; prioritising user adoption can limit profits in the short term. This is why, in Xu’s opinion, platforms need to strategically consider how to build their ecosystem. “For a platform to achieve that growth and profit objective, then they need to really think about network effects and their pricing decisions”, she thinks.

How academia collaborates with industry to accelerate growth

The basic precepts of network effects act as important background for the research questions that Xu is working on, using healthcare platforms as the setting. She’s collaborated with a leading health checkup platform in China to consider what the relationship is between network effects and different parties’ bargaining power within a platform. “[The healthcare] sector is huge in China”, Xu says. “On an annual basis, there are over 500 million checkups being conducted, with a market valuation of $30 billion.” The platform enables patients to book preventative healthcare services directly with providers; Xu notes that, because these services are paid out of pocket in China, the platform functions much like an e-commerce marketplace.

In its first five years, the platform experienced exponential user adoption on the patient side. “This is kind of the dream for most of the platforms out there”, Xu says. However, the supply side was constrained by the need for account managers to facilitate and maintain the relationships between platform and providers. Once Xu had access to the platform’s data she was able to develop a model that captured the decision-making factors for users on both sides of the platform. This gave her the basis for making recommendations, she tells us: “From there, we can give guidance to platforms to know where to allocate their marketing resources to either achieve this growth objective or achieve the profitability objective.”

By working with a platform, Xu had access to real-world data to ground her research and enable her to explore deeper questions. In return, the platform benefited from the results of her tailored research to adapt its business strategy. It’s a departure from previous notions of ivory-tower academia, but a successful one, in Xu’s experience. “Valuable insights can be acquired without the need of running experimentation in a lot of those collaborations”, she says. Ultimately, Xu’s hope is that the benefits of partnership between academia and industry are clear and convincing – a potential avenue for the next wave of platform research.

To go further

This panel with Joost Rietveld, Jean-Pierre Vergne and Xu Zhang was part of the Platform Leaders event organised by Launchworks & Co on the 9th of November 2023 at the Science Museum. Attended in person and online by hundreds of experts from all over the world, Platform Leaders provides an opportunity for entrepreneurs, academics, practitioners and policymakers to understand the issues and shape the debate. Stay tuned for more valuable insights from the latest conference, check out the full list of speakers and agenda, as well as videos and articles, and join the community to learn about upcoming events and more.