ICT Platform-Based Ecosystem Diffusion and the Need for a New Architectural Perspective
ICT platform-based ecosystems are now restructuring the ways that businesses create and deliver value across a broad range of markets and industries, not just the information-intensive sectors (Downes and Nunes 2013). According to Choudary (2015: 23), we are in the midst of a transformative shift in business design as business models move from “pipes” (linear one-sided businesses) to “platforms” (multi-sided ecosystems). Although the one-sided business model served as the dominant design throughout the capitalist industrial era, new trends are now emerging at an exponential rate due to Moore’s Law (Ismail et al. 2014) as more platform-based ecosystems are disrupting a broader range of sectors including media (newspapers, magazines, books, music and TV); financial services and insurance, travel and tourism, real estate and hotels, automobiles, health and many others.
The key drivers behind the increasing growth and pervasiveness of platform ecosystems have been new technological trends such as the rapid adoption of smartphones, 3G and 4G internet connectivity, apps, cloud computing services, software embeddedness and digitisation, the Internet-of-Things and big data (Deloitte Centre for the Edge 2013). The proliferation of smart phone adoption and the ubiquity of internet connectivity via 3G and 4G networks has made it possible for new platforms to engage with a vast consumer audience.
According to the Deloitte Centre for the Edge (2013: 9-10), the cost of computing power has decreased significantly from $222 per million transistors in 1992 to $0.06 per million transistors in 2012. This has, in turn, decreased the cost-performance of computational power. Secondly, the cost of data storage has decreased considerably from $569 per gigabyte of storage in 1992 to $0.03 per gigabyte in 2012. The decreasing cost performance of digital storage enables the creation of more and richer digital information. Thirdly, the cost of Internet bandwidth has also steadily decreased from $1,245 per 1000 megabits per second (Mbps) in 1999 to 423 per 1000 Mbps in 2012. The declining cost performance of bandwidth enables faster collection and transfer of data, facilitating richer connections and interactions. Additionally, the use of the Internet continues to increase creating widespread sharing of information as more people are now connected via mobile devices (Deloitte Centre for the Edge 2013: 9-10).
Apps and cloud computing services (software as a service, platform as a service and infrastructure as a service) have meant that entrepreneurs can scale new platforms very cheaply and very rapidly with minimal capital outlay i.e. Airbnb, Uber, Snapchat and Spotify (Downes and Nunes 2013). As more products have become Internet-enabled (the Internet- of-Things) with sensors or dematerialised through digitisation; and as many activities have been substituted by software robots; the rise and spread of platform ecosystems have increased. The data deluge created by these changes has also led to the emergence of platform firms with “Big Data” capabilities (using structured and unstructured data) such as Google, Amazon, Microsoft, Facebook and Alibaba who can perform high-speed predictive and prescriptive analytics (Sharda et al. 2014) which enables them to reduce costs, enhance their marketing and risk management capabilities and to outperform conventional one-sided businesses (Arthur 2014).
Although companies across industries are actively building platforms, these individual platforms are broadly different. For example, from the perspective of software developers, Android, Salesforce and Facebook Connect are vastly different. Medium and Wordpress are blogging platforms but have little in common with software development platforms. YouTube, Facebook, Instagram and Snapchat are described as social platforms, while Uber and Airbnb are referred to as marketplace platforms (Evans and Gawer 2016: 7). This becomes even more complex when one considers that the Nest Thermostat is called a platform and Nike is working on a platform to connect shoes, while GE claims to be using a platform approach to manage its factories (the Internet- of-Things).
The fact that these businesses are vastly different from each other creates problems when trying to plan strategies from two perspectives
- (Choudary 2015). First, how to plan strategy from the position of a newly evolving or established platform and second how to plan strategy from the position of an incumbent firm in an industry that is under the threat of disruption from a platform ecosystem i.e. Nokia’s recent demise at the hands of the Apple iPhone. Research undertaken by Choudary (2015), revealed that across all types of platform three distinct architectural layers repeatedly emerged. These three layers consisted of:
- 1) The network or marketplace community.
- 2) The infrastructure.
- 3) The data.
This has made it possible to formulate a unifying architectural framework - referred to as the “Platform Stack” (see Fig. 4.2) - to explain the different types of platform configuration. This forms an important basis from which future platform strategies can be planned. Each of these configurations will now be analysed in more detail starting with the network-marketplace community.
Network-Marketplace-Community-Layer: the first layer of the platform comprises participants and their relationships and includes social networks. This also involves the matching of buyers and sellers with regards to goods and services. Some platforms may have an implicit community layer. For example, users of Mint.com are not connected to each other but every user’s financial analytics are benchmarked against that of similar users. According to Choudary (2015), every user benefits implicitly from the community without the requirement to connect with others explicitly. So the external network of producers creates value in the network layer. However, to enable this value creation, platforms need a second layer: infrastructure.
Fig. 4.2 The platform stack (Adapted from Choudary 2015)
Infrastructure Layer: this layer encapsulates the tools, services and rules that enable interaction to take place, this is sometimes referred to as ‘plug- and-play’ (Choudary 2015). This layer has little value on its own unless users and partners create value on the platform. External producers build on top of this infrastructure. For example, on Android, developers produce apps, on YouTube video creators host videos and on eBay, sellers host product availability.
On development platforms such as Android, the infrastructure layer may be very dominant. On other platforms such as Instagram, the infrastructure layer may be thinner. Therefore, the infrastructure layer provides the infrastructure on top of which value can be created i.e. the software upon which application programmes can run or other services. However, large-scale value creation leads to the problem of abundance. With an abundance of production, search costs increase for consumers. Too many videos on YouTube may make it harder for consumers to make a selection. To solve this problem, the platform stack needs a third layer: data.
Data Layer: this is the final platform layer. Every platform uses data since the data helps the platform to match supply with demand. The data layer creates relevance and matches the most relevant content/goods/ services with the right users. In some cases, the data layer may play a very dominant role. For example, GEs Predix, Internet-of-Things (IOT) factory platform is data-intensive.
While platforms function across these three layers, the degree to which each one dominates may vary. The platform stack helps to reconcile the differences between different platforms while also acknowledging the similarity of the business models across all these instances (Choudary
2015). For understanding the different types of platforms, the chapter will now explore three basic configurations of the platform stack in more depth.
Basic Configuration 1 - The marketplace/community platform: Airbnb and Uber and most marketplace platforms have a thick marketplace/ community layer, and the network is the key source of value. Online communities like Reddit, social networks like Twitter and content platforms like YouTube benefit from thick or dense community layers. All three layers play a role although one may be more dominant than the others. The stack helps to illustrate that every platform will have its unique configuration. Certain platforms, like Craigslist and some online platforms, focus almost exclusively on the marketplace or community layer with almost no infrastructure and without much leveraging of data.
Basic Configuration 2 - The Infrastructure Platform: development platforms such as Android provide the infrastructure on top of which apps may be created. In tandem with the Google Play marketplace, Android’s development infrastructure is the key source of value for developers. Traditionally development platforms have focused on the infrastructure layer without a marketplace for apps. As a publishing platform, WordPress provides infrastructure exclusively. It doesn’t provide network benefits or any value through data.
Basic Configuration 3: The data platform: the third basic configuration is the one where the data layer plays a dominant role. The data layer plays an important role on every platform. Facebook uses data to fashion newsfeeds, and Airbnb uses data to match hosts to travellers. However, on certain platforms the data layer itself constitutes the key value created on the platform. Some of them may not even seem like platforms, but they follow the same stack while focusing almost exclusively on the data layer. Wearables are a good example, Nike’s shoes and Fuelband constantly stream data to an underlying platform that integrates the user experience across the shoe, the wearable and the mobile apps. Wearables such as Jawbone create value through the data platform. The wearable produces data constantly, and the platform provides analytics back to the user based on the data. The platform also pools data from many users to create network-level insights. Wearables, therefore, benefit from implicit network effects (Baldwin and Woodward 2009).
The Nest thermostat and the Internet of things are also good examples. The Nest thermostat uses a data platform to aggregate data from multiple thermostats. This aggregation of data enables analytics for thermostat users and powers services to the city’s utilities board. The Internet-of- Things (IOT) will also give rise to new business models in similar ways through the creation of data platforms.
Finally, GE is focusing on the “Industrial Internet” which is another example of a data platform. Machines embedded with sensors constantly stream activity data into a platform that helps each machine learn from other machines and provides network intelligence. These machines benefit from implicit network effects, and every machine learns from the community of machines it is concerned with.
If a platform is to scale successfully, it must be centred on the goal of value creation. In terms of the Platform Stack, this is known as the “core value unit” concept (Choudary 2015). The core value unit is the minimum standalone unit of value that is created on top of the platform. This will depend on a large extent on how the platform is configured. For example, the core value unit could be network/marketplace/community- dominated, infrastructure-dominated or data-dominated.
The core value unit on platforms that have a dominant network/market place/community will be the goods and services that they offer. Where the platform acts as the underlying infrastructure on top of which value is created then apps form the core value unit i.e. on development platforms. Meanwhile, the minimum unit of content constitutes the core value unit on a content platform i.e. videos on YouTube. Finally, on data-dominated platforms, the data itself is the source of value. For example, on a retail loyalty platform, the data profile of the consumer is the value unit. It is the core source of value to a retailer interested in targeting that consumer.
When implementing platform scale, successful platforms such as Uber, Airbnb, Facebook, YouTube and Upwork always start at the infrastructure layer first (Choudary 2015). It is important to build the infrastructure first in order to enable interactions to take place in the layer above. As the infrastructure gains adoption, an ecosystem of producers and consumers starts to evolve. For example, drivers and travellers start using Airbnb and developers, and users start adopting Android. This becomes the next discernible stage in the evolution of the platform. Finally, activity by producers and consumers on the platform generates significant amounts of data. The data layer then serves to make future interactions more efficient and keeps users regularly engaged in the platform. As the data layer grows stronger, the network or ecosystem layer also increases in strength.
Most multibillion dollar start-ups (Choudary 2015: 319) have achieved platform scale using this architecture (Amazon, Google, Facebook and Alibaba, etc.). However, although this template works for start-ups, it does not work for traditional one-sided businesses seeking to develop a platform. Traditional businesses according to Choudary (2015: 320), lack a culture of data acquisition and data management. Choudary (2015: 320) therefore recommended that the journey to platform scale needed to start with the data layer, followed by the infrastructure layer and then the development of the network-marketplace community. Choudary (2015) recommended five key stages in this evolutionary development:
- 1) Build a culture of data acquisition.
- 2) Enable data porosity and integration.
- 3) Leverage implicit data-driven network effects.
- 4) Build explicit communities.
- 5) Enable explicit exchange.
The first stage for a traditional business, according to Choudary (2015: 321), was to create a culture of data acquisition. The firm needed to understand that higher data acquisition meant greater monetisation opportunities. All digital services that are introduced to users should be integrated at the data layer, and every service should seek to acquire data that can be monetised in some form in the business. A strategy that intended to leverage platform scale should therefore, start with a coherent data strategy.
Once a strategy of data acquisition had been established, the second stage was to institute infrastructural change by integrating the internal organisation. According to Choudary (2015), the firm must integrate all processes, workflows and touchpoints at the data layer. Firms must restructure their internal systems to be more data-porous with internal application programming interfaces (APIs) and avoid silos that prevent cross-communication. The third stage is where the firm starts to leverage its existing user base. Once users have been profiled on the database the business can start to target them with recommendations etc. Once the first three stages are complete, the firm should then start to build a community. There has been a tendency (Choudary 2015: 324) for traditional firms to skip the first three steps and then fail because of the inability to leverage intelligence due to the lack of integration at the data layer. If the firm reaches the final stage, it will be able to operate as effectively as a modern platform company.