Does a meta-search platform favor its affiliated sales channels in its ranking?
This blog is written by Reinhold Kesler.
Digital platforms have become important intermediaries in many markets, be they traditional or newly created ones. Their key promise is to bring about lower search and distribution costs, better matches of market participants, and transparency about offers. However, they also have been shown to be able to steer consumers toward certain products and suppliers through their recommendations. This steering is increasingly often met with concern, especially with the presence of vertical integration of platforms along the consumer journey and possible incentives to bias the recommendation.
The corresponding policy debate is shaped by prominent cases of the European Commission that involve Google favoring its comparison shopping service over others and Amazon’s hybrid role as a marketplace and seller giving an edge over third party sellers. In this respect, the European Digital Markets Act (DMA), which entered into force in November 2022, aims to restrict the power of large online platforms that are designated as gatekeepers. These are subject to do’s and don’ts with regard to a range of business practices. One of the prohibitions comprises self-preferencing, which forbids treating own services and products more favorably than those of a third party.
In a recent study published in Quantitative Marketing and Economics, we explore self-preferencing in the context of online hotel booking. In particular, we empirically study whether an integrated meta-search platform favors its own affiliated sales channels. Meta-search platforms pool offers from different hotels – as do online travel agents like Booking.com and Expedia – but, in addition, for each hotel, they display the different sales channels available from which they predominantly retrieve payments through cost per click (CPC, see Figure 1, left). This gives a vertical and horizontal ranking, respectively (see Figure 1, right). The two rankings allow a price comparison on a more aggregate level, thereby making these platforms often the starting point of a consumer journey towards booking a hotel and economically relevant, according to a sector inquiry by the German competition authority. Interestingly, the two major online travel agents (Booking.com and Expedia) each own a meta-search platform (Kayak and Trivago), where the respective acquisitions raised concerns of search bias in favor of related sales channels.
We web-scraped search results for overnight stays in Paris from 2014 until 2017 on Kayak, the meta-search platform already belonging to the Booking Holdings at that time. In turn, we look at the determinants of both the vertical and horizontal rankings and, in particular, whether the company affiliation plays a role.
For the horizontal ranking, we find that online travel agents of the Booking Holdings are more often in the most prominent spot than they are among the cheapest sales channels. In regressional analyses accounting for differences in prices and popularity, among other factors, we indeed find Booking-affiliated online travel agents to be more likely among the visible and most prominent sales channels on Kayak than competing online travel agents. For the vertical ranking, the results suggest that hotels are ranked worse when rival online travel agents, i.e., the ones affiliated with the Expedia Group, are the cheapest sales channel. On average, such hotels are ranked eight positions worse. We do not find this pattern for the vertical ranking in an analogous empirical analysis on Google Hotels, a meta-search platform that is not vertically integrated with an online travel agent.
However, our empirical analysis is also subject to caveats. First, we are not able to observe the actual cost per click paid to Kayak by sales channels, which may vary and are presumably taken into account for the rankings. We try to address this by analysing the non-integrated meta-search platform Google Hotels, which potentially experiences a similar heterogeneity in payments and distinguishing chain and independent hotels, which may differ with respect to setting the CPC. Second, a concern may be the presence of systematic differences across sales channels unobserved to us (e.g., one channel having a breakfast option), which are potentially leading to better ranking positions. However, Kayak’s main goal is to provide comparability across offers, and Google Hotels may again serve as a benchmark, while this kind of differentiation of amenities is also not clear given similar commission rates across online travel agents.
Assuming these caveats to be less problematic, the results indicate that Kayak takes joint revenues of the integrated firm into account (i.e., both commissions and cost per click, see Figure 1) and that it favors affiliated sales channels.
Although we cannot provide a definite conclusion on a socially optimal ranking, there are potential risks of ranking optimization by a vertically integrated meta-search platform, as suggested by the results. First, such a ranking may diverge from consumer interests and, by this, lower search quality. Second, worse ranking positions that come along with lower prices elsewhere may work similarly to price parity clauses that have been abolished in some European countries and are also prohibited for gatekeepers designated under the DMA.
This brings us back to the policy debate revolving around the power of digital platforms and how to warrant a contestable and fair digital economy. Our article provides a case in point to study how vertical integration affects the recommendation of products and suppliers by a digital platform but also demonstrates the challenges of the empirical analyses which necessitate better data. More generally, such empirical studies can inform the crucial debate about the implementation and enforcement of regulations involving business practices like self-preferencing.
Cure, M., Hunold, M., Kesler, R., Laitenberger, U., & Larrieu, T. (2022). Vertical Integration of Platforms and Product Prominence. Quantitative Marketing and Economics, 20, 353–395.
It was another strong year for academic research on platform competition
In this post I take stock of the academic research on platform competition published in 2022. Using Platform Papers data, I look at the volume of research published across academic subdomains and how it compares to prior years. I further look at some of the themes covered by this research and how it maps on to developments in ‘the real world’. Finally, I briefly look ahead to 2023.
A look at the numbers
Management & Organizations: 39 papers
Information Systems: 24 papers
Marketing: 21 papers
Economics: 16 papers
The share of papers published in marketing journals this year stands out. While the disciplinary subsamples are fairly small and therefore volatile, it does seem that marketing scholars are increasingly interested in platform competition research: Whereas in 2018 papers published in marketing journals accounted for less than 3% of all papers added to the database, in 2022 this was 21%.
Management Science was by far the most prevalent outlet with 22 platform papers published in 2022 (of these, 13 were published in marketing departments). Other popular journals include Information Systems Research (12 papers) and the Strategic Management Journal (11 papers).
A look at the content
Thematically, there is a pretty even split between research covering issues related to ecosystem governance and orchestration (35 papers), research studying network effects, winner-take-all dynamics, and pricing (31 papers), and research addressing heterogeneity within and between platforms (31 papers). Corporate scope (e.g., platforms vertically integrating into the complementor space) received considerably less attention this year with 17 published papers.
Below I list five papers that got published in 2022 that I am particularly excited about:
Fending Off Critics of Platform Power with Differential Revenue Sharing: Doing Well by Doing Good? (Bhargava, Wang & Zhang; Management Science). I like this paper because it ties into the ongoing discussion about whether and how to regulate dominant platforms. The authors look at a very specific but oft-debated policy: dominant platforms changing their revenue sharing structure to be more favorable to smaller, less successful sellers. The economic model suggests that such a policy change not only benefits small sellers on the platform, but also larger sellers, and ultimately the platform itself. As the authors note: “Hence, an intervention that ostensibly offers concessions and generous treatment to producers might well be self-serving for platforms and good for the entire ecosystem.”
From proprietary to collective governance: How do platform participation strategies evolve? (O’Mahony & Karp; Strategic Management Journal). This paper tracks how a platform’s governance evolves over time and how it affects user participation on the platform. It’s another important contribution to illustrate that governance is highly dynamic and strategic. Platforms that open up attract higher participation rates, but participation by users declines when the platform’s framework of rules becomes unclear. (Twitter, anyone?!)
Positive Demand Spillover of Popular App Adoption: Implications for Platform Owners’ Management of Complements (Lee et al.; Information Systems Research). Traditional frameworks for analyzing competition do not unequivocally apply to platforms. Whereas competition from successful market participants can be detrimental in traditional markets, this paper finds strong empirical support for positive spill-over effects. Popular apps increase the adoption and usage of non-popular apps on the platform. These effects apply to apps released prior to the launch of the popular app as well as those released after.
Local Network Effects in the Adoption of a Digital Platform (Kim et al.; The Journal of Industrial Economics). Not all network effects are created equal and managers and academics are increasingly coming to grips with this. Studying the fantasy sports markets, this paper finds that “the size of a county’s existing user base on the platform significantly impacts the number of new adopters in that county, while the size of the user base in nearby counties does not.” In other words, participation by users that are more closely connected in the real world results in stronger network effects on the platform.
The Future of the Web? The Coordination and Early-Stage Growth of Decentralized Platforms (Hsieh & Vergne; Strategic Management Journal). This time last year some platform managers might have worried that decentralized platforms would soon put them out of work. The ‘Crypto Winter’ will have somewhat cooled those expectations, but decentralized governance of platforms certainly hasn’t gone away entirely either. This pioneering paper analyses 20 cryptocurrency platforms to better understand how decentralized platforms are governed. The authors document three distinct governance mechanisms: 1) algorithmic coordination, 2) social coordination, and 3) goal coordination.
Speaking of which, I am grateful for all the scholars who have contributed their time and knowledge by translating their platform competition research papers into digestible blogposts. While I highly recommend going back and reading all of these excellent blogs (as well as subscribe to ensure you don’t miss out on any future blogs), here are the three most read blogposts from 2022:
Big Tech Platforms’ Entry into Healthcare and Education (by Özalp, Ozcan & Gawer). This blog describes the process of ‘digital colonization’ where firms like Apple and Google enter highly regulated markets such as health care and education by deploying various forms of data capture to generate data-driven insights, ultimately translating into new products and services, while being mindful of the highly regulated and sensitive nature of these industries.
Platform Envelopment and Network Effects (by Allen, Chandrasekaran & Gretz). This blog describes how envelopment—the strategy of absorbing the core functionality of another, often complementary platform—can help new platforms reduce their dependence on network effects and solve the chicken-and-egg problem. The main idea is that by absorbing an outside technology, firms can increase the standalone value of their own platform.
When Freemium Succeeds (by Rietveld). In this blog, I discuss how social product features such as multiplayer modes for video games, ridesharing functionality in ride-hailing apps, and virtual collaboration tools in productivity software can both help and harm a freemium product’s widespread adoption in its respective market. Social features can generate network effects, but they may fall flat if the addressable market for a product is insufficient.
As regulation gets implemented and antitrust enforcement continues to target digital platforms, I am confident that 2023 will be another eventful year for platform competition research. My objective is to keep updating the references dashboard on a monthly basis and publish blogposts based on outstanding academic articles at similar intervals. I also hope to add some new functionality to the website and ensure it keeps running smoothly. Notably, I am part of the organizing team for the first summit of the European Digital Platform Research Network (EU-DPRN). This will be a two-day academic conference on digital platforms and ecosystems taking place on June 8 and 9th in Milan, Italy. The program will be accompanied by a one-day practitioner conference. More details to follow soon.
I hope you enjoy following Platform Papers and that you continue to do so in 2023. If you have any feedback or suggestions for improvements, feel free to reach out to me directly.
Happy New Year!
 The methodology for obtaining these inputs has remained the same. The methodology for obtaining papers is described in great detail in my platform competition review article in the Journal of Management.
 These numbers add up to more than 100 because some papers are assigned more than one theme.
Successful platform ecosystem orchestration requires more than network effects alone
This blog is written by Melissa A. Schilling.
One of the key jobs of a platform ecosystem’s sponsor (or “hub”) is to find a way to overcome the classic “chicken-and-egg” problem that new platform ecosystems typically face, i.e., the ecosystem may not be valuable to one or more types of participants until it has a critical mass in other types of participants. For example, video game consoles need games to be attractive to attract buyers, but game developers prefer to develop games for consoles that already have a lot of users; drivers are not interested in joining a ride sharing platform until there are a lot of customers, and customers get little value from using a ride sharing service until there are a lot of drivers, and so on. As a result, solving the “chicken-and-egg” problem has received considerable attention by both scholars and managers (see a great blog post here), and many platform managers see this as the crux of thjob.
Gettingthe point where the platform ecosystem has enough of the right kinds of participants is crucial to survival and should be celebrated, to be sure, however this is just where it starts getting interesting. A strong platform strategy should leverage several other levers to increase the value creation and capture in the ecosystem. I’ll discuss three here: Selective promotion to unlock stars and manage customer perception of depth and breadth, leveraging the data to build and refine new products and services, and facilitating scale benefits individual complementors could not achieve on their own.
Through selective promotions like endorsements, awards, marketing campaigns, and more, the platform hub can direct attention to high quality complements that deserve more attention than they are getting, thereby increasing the likelihood that they become “stars” in the ecosystem. This type of attention directing by the hub can be very powerful. Apple provides a great example: When Apple features particular applications on the home screens of its iOS App Store in categories like “Editor’s choice,” “App of the Day,” “Best new Games,” those applications may get up to six times as many downloads as other applications during the period they are featured.
Furthermore, through selectively targeting different types of applications, a platform hub can manage end-users’ perceptions of the range and overall quality of the ecosystem, and can spur consumers to try a broader range of products from the ecosystem. Consistent with this, research by Rietveld, Schilling and Bellavitis (2019) on video games found that platform sponsors select games for endorsement not only based on their quality and sales performance, but also on the degree to which they can unlock unrecognized value in the game, and the game’s potential to enhance the balance of the overall portfolio. Specifically, platform sponsors were more likely to endorse games that had high quality and good initial sales but were not market leaders. Additionally, they were more likely to endorse games that were in a high-value genre.
Leveraging the Data
The hub of a platform is often in a unique position to capture and utilize the data generated by the platform. Many platforms use the data to create better experiences for their customers, such as through providing recommendations and reviews. However, platform hub managers often do not fully leverage the opportunity to more proactively collect, use and sell data that could increase the value of existing complementors, or to catalyze the creation of complements that do not yet exist.
For example, lodging platforms like Expedia and Airbnb have access to exceptionally rich data on which lodging options customers choose and at what price. Not only do they have aggregate data on market trends in lodging, but they can also track a user’s choices over time and assemble a portfolio that tells them a lot about what that customer values and how much they’re willing to pay for it. They also have access to review scores and comments lodgers leave after their experiences. Both Expedia and Airbnb provide the review scores to other customers and utilize them (to varying degree) in search rankings, and Airbnb uses customer choice data in its dynamic pricing recommendations it makes to hosts. However, both platforms could be making considerably more use of their customer data. For example, both platforms could be providing data to their lodging providers on the degree to which customers would value additional features such as in-room dining, basic kitchen appliances, acceptance of pets, etc. They are uniquely positioned to help lodging providers in a given locale differentiate themselves to better tap underserved market segments. They can even identify geographic locations in which particular market segments are not being served at all and advise developers on opportunities that exist to develop or expand the range of hotel offerings, or they could even choose to develop these lodging options themselves!
One company that has done this exceptionally well is the popular movie streaming platform, Netflix. In 2017, Netflix started Netflix Studios, and began recruiting some of television’s most successful writers and producers to start making original content in house. By 2021, Netflix was spending over $5 billion on original content, making it one of the largest film production companies in the world. For a movie rental service to vertically integrate into developing its own content seemed a peculiar move at the time, because making films and television shows requires fundamentally different technology, equipment, personnel, and expertise than distributing films and television shows. What could a specialist in media distribution know about media production? A lot, it turns out.
Netflix’s rapidly growing datasets enabled it to know which customers liked which films, which genres were growing, which new stars were gaining followings, which new production houses were gaining traction and more. The relationships it had cultivated with small independent filmmakers and budding actors also helped ensure the firm’s access to a pipeline of new creative talent and helped build goodwill toward the company. Sean Fennessey, a writer for pop culture website The Ringer, explained how important Netflix was to frustrated filmmakers who could not raise enough support to get a major studio movie off the ground, “To the creators stifled by the rise of Hollywood’s all-or-nothing focus on franchise films, Netflix felt like salve on an open wound.”
Netflix also used its massive distribution reach and selective promotion to drive viewers to its original content, building audiences for its series and crafting its reputation as a first-tier production house. Netflix profited in multiple ways from its original content: Having popular exclusive shows helped attract and retain subscribers, and having both a large audience and a powerful library of original content gave it more bargaining power when negotiating license fees for content produced by others. Collectively, it was a powerful advantage.
In some platform ecosystems, the platform hub’s diversification into its complementor’s business could create more harm than benefits. Complementors with many platform choices, for example, might prefer opt to participate in platforms in which the hub is not a direct competitor – a dynamic known as “channel conflict.” In this situation, the platform hub can instead offer its data and advisory services to existing complementors and would-be complementors rather than entering these businesses through direct ownership.
Facilitating Scale Benefits
Another way in which a platform hub can help to unleash greater value in its ecosystem is through identifying those activities of complementors that would benefit by greater economies of scale, and either providing assets for those activities that complementors can access or providing another means by which the complementors can pool their scale. For example, if multiple complementors would benefit from the development of a powerful data analytics engine or sophisticated advertising capabilities, the platform hub can either provide those activities itself, or help to convene collaborative relationships that enable its complementors to share that effort and expense.
A great example of this is provided by Soteria Investments, a platform created to facilitate the buying and selling of distressed debt. The buyers and sellers of distressed debt are of highly variable size; a handful of large banks and investment firms make hundreds of transactions a year, while the vast majority of players make less than a dozen transactions a year. The buying and selling of distressed debt was historically a human-mediated transaction – sellers either searched for buyers directly among their contacts, or hired an investment banker to search on their behalf and then paid that investment banker a retainer fee, commission, or both. Furthermore, the deal flow in distressed debt is very segmented by geography and industry – a given seller might only have construction loans in the Midwest for sale, for example, giving them relatively little exposure to overall trends in deal flow and pricing, while also giving them inadequate incentive to invest in acquiring and analyzing a broader and deeper base of data. By providing a platform for buyers and sellers to find each other, Soteria helps a wider range of buyers be exposed to a wider range of sellers, while also collecting multinational and multi-industry data on deal flow and pricing. Access to that pooled data also gives it both means and incentive to invest in state-of-the-art data analytics capability that it can then offer to participants in its ecosystem, helping them to achieve more efficient pricing, greater control over risk, and faster transaction consummation.
Notably, facilitating the pooling of scale has another benefit to the platform hub: by enabling complementors to obtain the benefits of larger scale without actually achieving larger scale can help keep the complementors on a more level playing field, preventing one or a few from rising to a dominant position that increases their bargaining power over the hub.
Now you’re ready to orchestrate!
This process of managing an ecosystem to help participants be more successful both individually and in combination is usually termed “orchestration.” The platform manager is like a conductor that directs all the players to perform in ways that come together into a harmonious whole. It’s a complicated job – there are complex competitive dynamics and other interdependencies between the participants in an ecosystem that require careful thinking through. The platform hub manager must also take care to not be too heavy handed lest they alienate their complementors – rather than hierarchical authority, most platform hub managers rely on incentives and guidelines that are to some degree jointly negotiated with the complementors themselves. But if orchestration is well done, the ecosystem becomes much more powerful and valuable than the sum of its parts.
Rietveld, J., Schilling, M. A., & Bellavitis, C. (2019). Platform strategy: Managing ecosystem value through selective promotion of complements. Organization Science, 30(6), 1232-1251.
How brands can regain access to their consumers by building their own platforms
In today’s platform era, marketplaces and matchmakers such as Amazon, Alibaba, and co. have become powerful intermediaries to brand offerings: These brand aggregation platforms offer consumers virtually endless shelves, provide product recommendations and reviews, and allow them to easily compare offerings. As such, traditional brands struggle to keep direct contact with its consumers and to build and maintain customer relationships.
But brands are not standing by idly; they are building own platform offerings with the goal to reconquer the direct interface to consumers. Importantly, these platforms are not mere copies of the highly transaction-focused Amazon and its peers. Instead, they focus on increasing the number of touchpoints and establishing direct and intimate relationships with consumers. To do so, these brands build upon their core offering and provide value around it by holistically addressing consumers’ various needs associated with the core offering. Take Nike’s training platform Nike Run Club as an example. It offers an expansive array of functionalities revolving around Nike’s core offering of sports apparel such as a community of fellow runners, running competitions and challenges, and routing and tracking features. As a result, Nike can establish and leverage daily touchpoints with many consumers and extend its core offering to comprehensively address consumers’ running-related needs. We call this novel type of platform brand flagship platform, which we introduce and discuss in the paper “The Platformization of Brands” published in the Journal of Marketing as well as an accompanying article in the Harvard Business Review entitled “Building Your Own Brand Platform”. In the following, we analyse these nascent brand flagship platforms and offer guidelines for brands looking to platformize.
We conceptualize digital platforms as places where consumers and third parties crowdsource (crowdsend) products, services, and related content generated by (provided to) the platform participants. That is, rather than only the brand itself providing value, a brand flagship platform considerably expands its offering while simultaneously maintaining feasibility by incorporating consumers and third parties into the value creation process. Let´s return to the example of Nike Run Club for illustrative purposes. The platform does not only sell custom sportswear but also offers social events, fitness challenges and leaderboards created by the community. Moreover, users can find expert guidance, motivational music playlists, personal training by third-parties as well as exclusive Nike products on the platform. Many of these functionalities only thrive through the integration of consumers (e.g., challenges, leaderboards, and other community feature) and third parties (e.g., expert guidance, music playlists, coaching). Hence, consumers are benefiting from Nike´s expansion as they combine various elements of the service to find the most compatible product-service-content combination for their needs. Typically, the value created on brand flagship platforms can be classified by five distinct platform building blocks: transaction block (e.g., providing products for sale), community block (e.g., forums and profiles), benchmarking block (e.g., tracking runs, leaderboards), guidance block (e.g., AI coaching), and inspiration block (e.g., video content, blogs). Hence, brand flagship platforms do not only nor primarily serve as a sales channel for the brand. Instead their goal is to build an ecosystem surrounding the brand’s core offering by providing a mix of customized products, services, and content; and involving consumers, professionals, and third-party companies in the value creation.
The degree of crowdsourcing and -sending taking place on a flagship platform has important strategic implications for a brand because it affects how to set up and run the platform as well as which type of consumer relationship emerges from it. Hence, brands need to pursue the flagship platform style that best suits their strategic goals.
The platform as an instrument. These platforms enable low degrees of crowdsourcing and crowdsending so that value creation is strongly driven by the brand itself. As such value creation is typically focused on the brand’s core offering and the platforms’ functionalities tend to focus on enabling transactions or providing functional extensions to the brand’s core offering. As a result, relationships with consumers tend to be superficial as they use the platform only to meet immediate consumption-related needs. Example: Philips Sonicare
The platform as a guide. Platforms in this category foster high degrees of crowdsourcing but low degrees of crowdsending. Users of these platforms profit from the brand and third-party businesses but they usually do not create any significant value themselves. The goal of the platform is to expand the platform’s scope by inviting companies and professionals from related areas to join. By doing so, consumers are able to satisfy different but related needs which creates a genuine bond with the platform’s brand. Example: Whistle (by Mars)
The platform as a canva. Here, the brand encourages a high level of crowdsending while crowdsourcing is limited. As such, consumers can freely create value and interact with each other with limited oversight by the brand which requires a more laid-back platform management style. The high level of engagement with the brand leads to a strong attachment meaning that consumers strongly identify with the brand. At the same time, however, the limited oversight bears risks such as anti-social and fraudulent behaviors taking place on the platform. Example: Lego Ideas
The platform as a companion. These platforms enable both intensive crowdsourcing and -sending. As such, consumers are highly involved in the platform receiving and generating value. For example, a user that uses Nike Run Clun will benefit from the routes, challenges, and advice provided by other platform participants. But at the same time, that user will also provide value to the platform by uploading her runs, her favorite routes, or by providing own advice. Through this continuous give and take, this platform type is highly flexible and adaptable so that it is able to grow with the consumer. This is why Nike Run Club can remain relevant for consumers even as they advance in their skill and progress from total novice to avid runner. This perpetual companionship can elicit strong feelings of attachment and identification in consumers. Example: Nike Run Club; Bosch DIY & Garden
In order to do build a brand flagship platform, managers must rethink their market and start shifting the place of value creation from selling goods to satisfying consumers’ varied needs associated with the brand’s core offering. Given this substantial shift, the transition to a flagship platform should be a slow adaption process rather than an “all or nothing” approach”.
The transition consists of three important steps:
First brands should have a clear vision of the territory they want to cover. They determine how far and how fast they want to stray from their traditional territory while consumers define the needs that they want to see satisfied.
Second, each brand must decide which strategic goal(s) they are pursuing and adjust their platform offering accordingly. This also implies an adjustment of observed KPIs. Especially as a brand strives towards strong relationships, its KPIs evolve beyond sales towards a broader system that captures the engagement and value provided by and generated from the various platform participants.
Third, managers need to decide which parts of the value creation they want to make, buy (from third-party firms), or earn (from consumers). Importantly, a flagship platform with varied functionalities can only be feasible if the brand relegates some parts of the value creation to third-parties and consumers. Hence, managers must confidently give up some control in the process.
Wichmann, J. R., Wiegand, N., & Reinartz, W. J. (2022). The platformization of brands. Journal of Marketing, 86(1), 109-131.
Wichmann, J. R., Wiegand, N., & Reinartz, W. J. (2022). Building your own brand platform. Harvard Business Review, 100(5).
Absorbing functionality from another platform can help platforms solve the “chicken-and-egg” problem
Would you purchase a smartphone if it only offers a few compatible apps? Would you purchase a video game console if only a few games were available for the console? Is your answer ‘Probably not’?
Both these examples highlight an important relationship that characterizes hardware/software platform markets: The value of the platform (e.g., smart phone, video game console) increases as a larger amount of compatible software (e.g., mobile apps, video games) becomes available, also known as indirect network effects. Indirect network effects create a critical feedback loop: A platform with a greater amount of compatible software is attractive to consumers; more software becomes available as more consumers buy the platform.
This feedback loop creates a major challenge for new platform entrants and platforms with small networks in industries where these network effects are strong. It’s a classic chicken-and-egg problem: How can a platform grow their market if they lack sufficient quantities of related software? How can they attract software developers if they do not have enough adopters? Platforms in this position are dependent on software providers and possess low negotiating power.
So, what options does a new platform have in this scenario? It can try to build a stronger network, say by investing heavily in software provision (e.g., Nintendo making their own games). Or it can try to innovate radically to leapfrog the established competing platforms, which may be risky and difficult.
We propose a third strategy to compete effectively in these networked markets. In a new paper in the Journal of Product Innovation Management, we propose the use of platform envelopment in mitigating these indirect network effects. Platform envelopment is an innovation strategy, originally proposed by Eisenmann, Parker, and Van Alstyne (2011), wherein an existing platform absorbs the core technical features from another platform such that the functionalities of multiple platforms are combined. Envelopment involves more than just adding additional features; it is characterized by absorbing the core functionality of another platform. For example, a video game console absorbing the functionality to play DVD movies. We propose that platform envelopment can increase the stand-alone value of the platform, which should encourage consumer adoption even with limited software provision. This, in turn, decreases the platform’s reliance on software provision. Interestingly, platform envelopment has, thus far, largely been considered as a strategy for moving into an adjacent market. As a result, existing research has focused on how envelopment impacts performance in the enveloped industry (e.g., how a video game console offering DVD compatibility impacts the platform’s performance in the DVD market). In our paper, we look at the impact of platform envelopment on platform success in its original industry (e.g., how a video game console offering DVD compatibility impacts the platform’s performance in the gaming industry).
Platform envelopment offers two benefits to the platform. It makes the platform more attractive to the consumer (e.g., a video game console becomes more attractive with the ability to stream Netflix, Hulu, etc.). Critically, we propose that with envelopment, the amount of software becomes less important to overall platform performance. This is because envelopment attracts different segments of consumers who derive less value from the core functionality the hardware provides. Again, consider the example of video game consoles – consoles that only play games attract core gamers – all they care about are games. Core gamers value the console based on the available games. However, there is another segment out there – let’s call them novice gamers –who enjoy gaming but are on the fence of whether to buy the console because they are not sure if their novice interest is worth the investment. However, if the platform also doubles as a Blu-ray player, this might justify the purchase by novice gamers. These consumers, compared to core gamers, are not as influenced by the quantity of games available. Then, you have non-gamers who may purchase the console simply to use the Blu-ray player functionality; game provision is completely irrelevant for non-gamers. Importantly, our theory suggests that the net importance of software for the industry as a whole will decrease once these new consumers enter the market. Further, as the focal platform (e.g., video game console) is “borrowing” the network of the enveloped industry (e.g., Blu-rays), the marginal value of any one software supply is diminished. We call this phenomenon of weakening indirect network effects decreasing indirect network sensitivity.
We test our theory by examining the impact of platform envelopment on platform success in the video game industry with two studies. The first study looks at the relationship between hardware and software demand in the video game console industry in the US over several years and spanning several console generations. In the second study, we leverage data from several countries and consider this relationship in the context of a broad set of gaming platforms including gaming consoles, handheld devices (e.g., Gameboy), mobile phone gaming, online gaming, and PC gaming. Both studies find that platform envelopment increases the demand for the platform while simultaneously decreasing the importance of software supply on platform demand. Our findings suggest that a platform envelopment strategy can provide new entrants and firms with smaller networks a way to decrease their dependence on available software while increasing the platform’s stand-alone value, allowing them to better compete with larger platforms with established networks.
Can platform envelopment be an effective strategy in other industries? Let us consider video streaming. Compare the differences in strategy between Netflix and Amazon’s entertainment arm. Netflix relies mostly on third-party video content and is currently facing increasing costs for streaming rights. As Netflix offers no other functionality beyond streaming video content, it is dependent on content providers, which in turn leads it to raise the price charged to consumers. Netflix is now investing heavily in developing its own content to mitigate its reliance on movie providers. However, our research suggests that it could be beneficial for Netflix to consider streaming other content, such as music. Amazon’s entertainment arm utilizes this strategy by streaming both shows and music, making it less dependent on only one type of content provider.
Allen, B. J., Chandrasekaran, D., & Gretz, R. T. (2021). How can platforms decrease their dependence on traditional indirect network effects? Innovating using platform envelopment. Journal of Product Innovation Management, 38(5), 497-521.
Adding social features may help freemium apps become a superstar, but it could also backfire…
This blog is written by Joost Rietveld
Back in 2012, Sarah Needleman published an article in The Wall Street Journal titled “When Freemium Fails”. The article remarked that, despite the many attractive features of the popular digital business model, many freemium products fail to attract sufficient users, let alone generate revenues. Indeed, it is not uncommon for freemium products to have conversion rates (the share of paying users among a product’s overall user base) between two and five percent. Academic research, including by myself, later substantiated these observations: Freemium may offer users an attractive entry point into a new product, average use rates and customer spending on these products, however, tend to be well below that of traditional paid products.
Still, some of the most popular digital products nowadays are brought to market with a freemium business model. Examples include the hit video game Fortnite, the online dating application Tinder, and the file-hosting service Dropbox. Base versions of these products are offered at no cost to consumers, who can pay for optional content and features (such as character upgrades, exposure to romantic partners, or extra storage capacity) via in-app purchases. These examples represent a select group of freemium superstars—freemium products that are among the most-downloaded products in a market, which also generate substantial revenues for their commercializing firms.
This raises the following questions: Which freemium products are more likely to become a superstar? How do these dynamics differ from paid products?
The double-edged sword of social product features
In a recent study published in the Strategic Management Journal (open access), co-authored with Joe Ploog, we address these questions by focusing on a common design strategy for freemium products: the inclusion of social product features, such as multiplayer modes in video games (e.g., Fortnite’s Battle Royale), ride-sharing functionality in ride-hailing apps (e.g., Uber’s Pool), and virtual collaboration tools in productivity software (e.g., Google Sheets’ Share). Such social features enable interactions among a product’s user base, which, when present, enhance the product’s value in use.
Analyzing a large dataset of 9,700 digital PC games released on the Steam platform, we find that freemium products that incorporate many social features (such as online competitive play or local cooperative play) are 49 percentage points more likely than freemium products without any social features to become a superstar when they are released when the platform itself has a larger rather than smaller installed user base. Conversely, when the platform’s installed user base is small, freemium products are 26 percentage points less likely to become a superstar if they incorporate many social features. Notably, we find that these mixed effects do not apply to paid products.
Social product features can help freemium products attain widespread diffusion. By their very nature, freemium products enjoy strong social referral. We (as consumers) tend to be more likely to recommend products when they exhibit low risks to adoption (i.e., when they are free). We are also likely to share freemium products because we want to reciprocate to the products’ commercializing firm the benefits we receive for free. Social product features can amplify these tendencies. When the value of a product depends on the number of users, we might be even more likely to recommend a product, in the hopes of growing its user base and expanding the social engagement.
That said, a product with many social features likely won’t generate much value if it lacks a large user base. Indeed, a game like Fortnite isn’t much fun if there is no one to play with. Google Sheets doesn’t differentiate itself much from Excel if it wasn’t for the ability to collaborate in an online environment. Uber’s Pool functionality won’t help you save (on the environment) if there’s no one to share a ride with. Even worse, absent a large user base, consumers may feel they are missing out on key features if the product strongly relies on social functionality for its value proposition.
This is where the size of the platform’s installed base comes in. The platform’s installed base is a strong indicator of a product’s demand potential. Products released on a digital platform can only be adopted and used by those users who have first adopted the platform itself. When a product’s demand potential is constrained—because it is released on a platform with a small installed base—social engagement is less likely to occur. Users may be reluctant to adopt—let alone recommend—these products. To be sure, a freemium product without social features faces an equally constrained demand potential, however, it doesn’t require a large user base to confer its core benefits.
Finally, why, then, did we find that these mixed effects of incorporating social features do not apply to paid products? First, paying users are more likely to make an independent value assessment that is less reliant on social referral. After all, the decision to adopt a paid product is more consequential given the upfront payment that’s required. Second, users will, on average, spend significantly more time on paid products because they want to “get their money’s worth”. Put differently, users may dip in and out of freemium products because they do not incur any costs to adopt these products in the first place. Combined, these differences suggest that, even when paid products incorporate social features, they will be less reliant on widespread diffusion to create value for their users.
Freemium is a difficult business model to get right. Despite its many appealing features, most freemium products fail. That said, when freemium products do succeed, they tend to succeed to the point where they become superstars. These superstars enjoy disproportionately more users, more usage, and more in-app purchases. Our research on digital PC games shows that freemium products are more likely to succeed by incorporating social features when they are launched on a platform with a large installed base, whereas freemium products without any social features are more likely to succeed when they are launched on a platform with a small installed base. These findings hold important implications for firms’ product design choices and for their product-market strategies.
Rietveld, J., & Ploog, J. N. (2022). On top of the game? The double‐edged sword of incorporating social features into freemium products. Strategic Management Journal, 43(6), 1182-1207.
How building trust in online platforms can help users but may harm intermediaries
Intermediaries are everywhere in our economy: brokers in the finance and insurance industries, headhunters in the labor market, distributors in retail, housing agents in real estate, and online platforms in the information technology industry, just to name a few. In 2010, intermediaries contributed an estimated 34% of the US gross domestic product. However, all intermediaries face the risk of disintermediation, in which two sides circumvent the intermediary to transact directly and avoid the intermediary’s fees.
Disintermediation, wherein network members bypass a hub and connect directly, can be a big problem for any platform that captures value directly from matching or by facilitating transactions. Imagine that you hire a plumber from a platform like Homejoy and are satisfied with the service. Would you really go back to Homejoy to hire the same plumber again? After a consumer is matched with a service provider, there’s little incentive to return to the platform. Additionally, after obtaining enough clients from a platform to fill his or her schedule, the service provider won’t need that platform anymore. This was exactly the problem that doomed Homejoy, which shut down in 2015, five years after it was founded.
Examples like this are not hard to find. For instance, the traditional role of book publishers as intermediaries was weakened when Amazon enabled authors to sell directly to readers through its self-publishing services. Li & Fung, a supply-chain management company that connects global retail brands with Chinese manufacturers, suffered ongoing decline in revenue as retailers disintermediated to work with manufacturers directly. A survey by ZBJ.com, the largest online freelance marketplace in China, indicates that approximately 90% of transactions are conducted outside the platform after clients and freelancers have been matched on its platform. Hotels and airlines offer incentives to lure customers to book directly with them, thereby shrinking the revenue for online travel agencies.
How to address the issue of disintermediation? Because of the importance of disintermediation with regard to firms’ strategies and survival, platforms have used various mechanisms to deter disintermediation. Some platforms try to avoid disintermediation by enhancing the value of conducting business on them. But those services become less valuable once trust develops among platform users—and these strategies can backfire as the need for the platform decreases. This is what our study investigates.
In our recent study published in Management Science, we are curious to learn how user trust affects disintermediation, leveraging a randomized control trial (RCT) in a large online outsourcing platform. This platform enables clients to find freelancers who satisfy the clients’ job requirements to initiate and complete job contracts, and the platform charges a per-transaction service fee that is approximately 10% of each transaction’s value. In an RCT in 2015 which lasted for four weeks, the platform provider shows freelancers’ satisfaction scores (SSs) to a random sample of clients, illustrated in the Figure below. SSs are a newly developed measure of a freelancer’s business reputation based on his or her complete work history on the platform.
Our analysis of the above RCT generates interesting findings: As the platform improved its reputation-rating system, trust between clients and freelancers grew stronger, and disintermediation became more frequent, offsetting the revenue gains from better matching. We also identify other contributing factors to disintermediation tendencies, like being geographically near one another, having easily divisible jobs, and clients themselves having high ratings.
To platform managers, our findings highlight that disintermediation can sometimes render less effective an intermediary’s strategy to improve its profitability through enhancing trust. It is important to note that the main objective of our research is not to conduct a net benefit analysis to determine whether it is optimal for platforms to implement better reputation systems. Rather, because trust building is important for platform growth, it is of vital importance for a platform to build as much trust as possible. At the same time, our research suggests that as a platform builds more trust to facilitate transactions in its marketplace, it needs to adopt appropriate strategies to counter increased disintermediation.
Platforms could use a variety of strategies to reduce disintermediation as they enhance trust. Airbnb, for example, enhances trust and safety through host ID verification and background checks. At the same time, Airbnb reduces disintermediation by withholding host data, such as listing address or phone number, until the payment is made. Thumbtack, a marketplace that connects consumers with local service providers such as house cleaners, captures value pre-transaction: when customers post job requests on Thumbtack, service providers can send quotes to the customers; service providers pay fees to Thumbtack only if customers respond. Disintermediation affects Thumbtack less strongly because its model captures value before two parties agree to work together.
Other platforms recognize that the motivation to disintermediate comes from the service fees they charge and adopt different value-capture strategies to prevent disintermediation while still enhancing trust. For example, Chinese outsourcing marketplace ZBJ, which launched in 2006 with a 20% commission model, began pursuing other revenue sources after calculating that it could lose as much as 90% of its business to disintermediation. In 2014, ZBJ leveraged big data analytics to find that new business owners often used ZBJ to outsource logo design. However, after logo design, many of these clients would also need business and trademark registration. Thus, ZBJ began offering this service and has now become the largest provider of trademark registration in China. Replicating this experience, ZBJ began providing several other services to its marketplace participants. With these revenue streams, the company decided to significantly reduce its commission to 2% and shifted its resources from fighting disintermediation to growing its user base and building trust (for example, by encouraging clients and freelancers to communicate). Because of these changes, the company obtained a valuation over $1.5 billion in 2018.
Platform managers may use our study as a starting point to come up with more creative strategies to mitigate the tension between trust building and disintermediation. We hope that our study inspires a growing stream of academic research as well as managerial best practices to prevent disintermediation and enhance platform value capture.
Gu, G., & Zhu, F. (2021). Trust and disintermediation: Evidence from an online freelance marketplace. Management Science, 67(2), 794-807.
 Spulber, Daniel F. 2011. “Should Business Method Inventions be Patentable?” Journal of Legal Analysis, 3(1): 265–340.
How Google and Apple use “digital colonization” to enter highly regulated industries
Big Tech platforms (GAFAM in the Western World: Google/Alphabet, Amazon, Facebook/Meta, Apple, and Microsoft), have risen to become the most influential and largest firms in the world. Their rise through their data-centric approach has been related to increased efficiency and innovations in many areas. However, there are also many concerns related to market power, legacy incumbents, and privacy which are now shaping specific regulations for these firms, as seen in the recent adoption of the Digital Markets Act (DMA) and Digital Services Act (DSA) by the European Parliament, and similar regulations that are upcoming in US, UK, as well as many other countries.
Although these regulations aim to broadly regulate the market power of platforms, there are some industries that are just now experiencing increasingly more activity by these Big Tech platforms and are witnessing initial signs of disruption: These are highly regulated industries such as healthcare and education.
These highly regulated industries are represented by their higher entry barriers, slower-moving nature, considerable state-involvement, and many incumbents protected from competition for a long time. As such, these industries are also associated with large potential for increased efficiency and innovation. This in turn makes the benefits and (potential) harms of Big Tech platforms’ activities critical in these industries since value creation is extremely important (e.g., patient lives saved by new technologies), yet concerns around data privacy, fair competition, and public services that are based on (human) rights are even more salient (e.g., medical or learning records already used by Google and others or “right to education”).
In a recent paper in the California Management Review, we explore how Big Tech platforms enter and compete in highly regulated industries. Our findings highlight that the crux of platform entry into highly regulated industries is access to sensitive data. Big Tech firms rarely aim to directly offer the “primary service” (e.g., providing school education or becoming primary healthcare providers) in these industries, but they rather focus on capturing data as a pathway to other value-adding activities. They do so by engaging two modes of data capture simultaneously: using their own hardware or software to build their own datasets (“direct data capture”, such as the continuous health data collected through an Apple Watch); and forming partnerships with state or private actors (particularly primary service providers) in the industry for access to existing data (“indirect data capture”, such as the UK’s National Health System’s (NHS) partnership with Google’s Deepmind where they provided patient data to Google).
As Big Tech firms combine the data they captured directly and indirectly, they can provide superior data-driven insights, which can add significant value to incumbent service providers (e.g., through saved lives, better learning outcomes, and lower costs). Ultimately, they may design and commercialize new products and services for the highly regulated industry, through which Big Tech platforms can capture value in multiple modes, including leveraging captured data in their other activities (e.g., advertising).
Overall, in highly regulated industries where access to primary service data (e.g., individual-level clinical data or educational records) is a bottleneck due to the sensitivity of this data, those firms that overcome this bottleneck, generally through indirect data capture combined with data-driven insights, gain a unique competitive advantage in the design of new products and services where incumbents (e.g., medical device manufacturers, textbook publishers) have historically operated without such data capture or analysis capabilities. This allows Big Tech platforms to change power dynamics in these highly regulated industries by providing unique added value through data-driven products and services that other industry players increasingly rely on, which in turn commoditizes these players over time.
We name this process (summarized in the figure below) of entry by Big Tech platforms in highly regulated industries “digital colonization”, which we specify as composed of four stages: (1) Provision of data infrastructure services to incumbents; (2) direct and indirect data capture in industry; (3) provision of data-driven insights; and (4) design and commercialization of new products and services.
For platform strategy, we suggest that superior data analytics capabilities that enable the generation of data-driven insights matter even more in highly regulated industries, as they offer a pathway to break the high entry barriers in these industries and solve the bottleneck of data access. We also find that, rather than replacing existing actors and competing head-on to offer primary services, successful platforms add value to the existing value chain in highly regulated industries through data infrastructure services, data-driven insights, and finally new products and services through a multi-stage strategy. Initially, platform managers need to negotiate access to primary service data. At the same time, they need to make efforts to capture novel data through proprietary hardware and software user interfaces. In this context, platform firms may find that subsidizing hardware and access to services can work effectively to “buy their way” into data access. After this initial stage, they can focus on capturing value through a variety of data-related industry activities, such as selling data-driven insights or designing new products and services.
Our findings also highlight the paramount importance of a platform’s policies and special procedures in dealing with the usually sensitive data within highly regulated industries. While taking precautions in the use of sensitive data may seem limiting for value creation (and capture) for a platform, this is necessary, as eventually, a platform’s value for its users is also driven by how well it balances its diverse stakeholders’ interests, especially in terms of data privacy and security, but also in terms of the explainability and fairness of its AI-driven activities.
Another important set of implications is for incumbent firms, who need to respond to Big Tech firms’ uniquely powerful form of competition. We suggest that incumbent firm managers need to formulate their own data capture and analysis strategies and decide quickly whether they will compete or partner with entering platforms. Eventually, incumbent firms need to consider the long-term implications of initially value-creating actions of Big Tech firms, such as providing data driven insights, and whether these will, over time, lead to the commoditization of incumbents’ own activities.
From a regulation and policy point of view, our study highlights that special consideration should be given to regulating access to data and the usage of technology providers in highly regulated industries. This is in line with recent work that suggests platform regulation should consider issues like “sharing (in situ) platform data” and “data mobility/portability” above and beyond a purely market power-based approach that is utilized in the utilities sector. A particular balance needs to be established between giving enough space for platforms to bring their data-driven innovations to various industries and setting clear guidelines as to what data they can access, use, and combine, and what additional responsibilities they must carry while operating in highly regulated industries.
Finally, considering the balance between value creation and capture, we highlight an important dilemma for policymakers and regulators: platforms with more advanced data capture and analysis capabilities can provide more significant added value through better prediction but also these are the very platforms that pose additional privacy and security concerns in a highly regulated industry with sensitive personal data.
To close off, given the dire need of innovation and increasing efficiency of highly regulated industries such as healthcare and education, finding ways to combine the benefits brought by digital platforms with rights, ethics, fairness, and respectful consideration of personal data will become an increasingly important problem to solve in the years to come.
Ozalp, H., Ozcan, P., Dinckol, D., Zachariadis, M., & Gawer, A. (2022). “Digital Colonization” of Highly Regulated Industries: An Analysis of Big Tech Platforms’ Entry into Health Care and Education. California Management Review, OnlineFirst.
Rural entrepreneurs exhibit worse responses to algorithmic changes, negatively impacting the efficacy of platform governance.
In many countries, rural areas have long been experiencing a troubling pattern of brain drain, leading to a dearth of business and employment opportunities. Meanwhile, in some parts of the world, the rise of digital platforms has created a path for rural entrepreneurs to earn a living through online selling.
It is one thing for entrepreneurs, whether urban or rural, to create and operate an online store, as some platforms have made it relatively easy to manage an online store – even by using just a smartphone. One troublesome aspect of online selling is that platforms frequently tweak their algorithms. This requires entrepreneurs to keep up with these changes and constantly adjust their strategy. If not, their products may not surface in customer searches.
The next problem is that platforms do not always communicate the changes in the most straightforward manner, leading to confusion among entrepreneurs. As we show in a recent paper, compared to rural entrepreneurs, urban entrepreneurs have access to offline sources of information that helps them “see through the veil”. This advantage can lead to lasting gaps in performance, especially after a major algorithmic change.
The Offline Interface for Online Platforms
For online business and digital platforms, our research highlights the importance to understand the offline interface – a term we use to describe the local, offline factors (be it economic, social, cultural or political) that affect participants’ ability to navigate a digital environment. This concept can be illustrated using the example from our paper, which explores how entrepreneurs respond to a major algorithmic change on a prominent e-commerce platform in China.
On e-commerce platforms, entrepreneurs usually start out as scavengers, either making their own products or pulling excess supplies from their environment (e.g., local factories). You see this phenomenon on platforms in both developing and developed countries. Entrepreneurs sourcing from nearby sources might sell socks, car radios and baby products, whatever assortment of products they manage to line up – or figure out what might sell based on a recent fad.
In May 2013, the platform examined in our paper was trying to increase the professionalism of its sellers (entrepreneurs) by encouraging them to specialize in a single product category. The goal was simple from the platform’s perspective: Over time, entrepreneurs would become category experts able to provide top-level customer service. To achieve this, the platform modified its algorithm to prioritize entrepreneurs who sold many products within one or two product categories (a characteristic termed “category focus”).
However, the platform had difficulty conveying this change to entrepreneurs. A somewhat cryptic announcement said that Big Data would be used to construct buyer profiles and that a new algorithm would “help entrepreneurs lock onto potential buyers and implement targeted marketing”.
Vague communication such as this is common. While hiring better communicators might help, the reality is that transferring complex information to others in a digital environment is more difficult than it seems. First, people do not always check their inboxes for platform-initiated communication. Email announcements are the poor cousins of the communications family, as they tend to be text-only. They may be crafted in a language that is poorly understood by much of the intended audience. If recipients see questions that need to be clarified, they can neither request nor receive immediate feedback.
As media richness theory has shown since the 1980s, communications that are richer in cues and more personalised, such as in-person discussions, are significantly more effective at helping decode potentially confusing messages. If we were to tell you about our research findings, we would probably do a better job sitting down and talking to you face-to-face than via this article alone.
Another reason for unclear communication is that a platform may want to hide the details of an algorithmic change – from its competitors, from public scrutiny and from participants themselves (to prevent some from “gaming the rules”).
These communication issues lead to equivocality in entrepreneurs’ interpretation of platform algorithms, and that equivocality is unevenly distributed across different segments of the entrepreneur population.
Rural entrepreneurs are left to figure things out by themselves
Let’s go back to the Chinese e-commerce platform. Based on our study of 2,395 randomly selected e-commerce entrepreneurs, the sales of rural and urban entrepreneurs were on par before the algorithm change. In fact, rural entrepreneurs, which comprised nearly 40 percent of the sample, sold a little bit more. But after the tweak, rural entrepreneurs actually went in the wrong direction by decreasing their category focus. This sank their stores’ rankings and resulted in sales that were 24 percent lower than those of their urban counterparts.
After four months, rural entrepreneurs had caught up in terms of category focus, but their sales remained inferior until several months later. This suggests that algorithms rewarded the entrepreneurs who “solved the puzzle” early, and that advantage was sustained over time.
Among urban entrepreneurs, guess who were the best performers? Those nearest to the platform’s headquarters. Their category focus doubled after the algorithm change. Moreover, it is telling that the biggest predictor of an entrepreneur’s behavior was not what their same-category competitors did, but what other entrepreneurs in their physical vicinity did.
In an online environment filled with irrelevant, vague and even false information, entrepreneurs can often be led astray. In such situations, the offline interface is all the more relevant: Knowledgeable sources of information in the local environment can help entrepreneurs address points of confusion and devise the appropriate strategy. Rural entrepreneurs are not less capable than their urban counterparts, as shown by their performance before the algorithmic change. Rather, their local, offline networks just don’t give them access to the same level of help to navigate the ever-changing platform environment.
What are examples of high-quality information sources? Our field research in China showed that discussions with experts – often former platform employees or individuals connected to platform designers – were important for entrepreneurs’ success. Self-organized learning groups are also more prevalent in cities. Yet, these opportunities to acquire valuable information were less available in rural areas.
We also met rural entrepreneurs who would take a long-distance bus to cities in order to learn more about the latest platform direction. As a rural entrepreneur told us: “Oftentimes, many rule changes on the platform are quite subtle and hard to interpret. But there are these platform gurus in Beijing, who can see through the veil. You must go on a ‘pilgrimage’ to obtain the ‘holy scriptures’ from [them].”
Lessons for platform governance
The first order of inequality in the digital economy is internet access. A number of initiatives are trying to fix this inequality, for instance the Next Billion Users projects by Google. But if we look at third-party entrepreneurs on an e-commerce platform, they theoretically have the same ability to reach all customers. Geography is not supposed to matter. Except that it does.
In practice, information asymmetry due to offline differences is common. This poses a major issue for platform governance, which is not unlike what governments face when implementing new policies or regulations.
Our research suggests that, before implementing major algorithmic changes, a platform needs to study its audiences carefully and assess their diverse information needs. Participants in remote geographic areas need special attention. This could mean creating learning groups with experts from nearby cities or sending clarifying messages or videos in a targeted manner, in which case the platform needs to implement well-designed pre-testing. The platform could also help information-disadvantaged participants by creating ranking algorithms that do not rush to punish those who fall behind on day one. In other words, algorithms that are less dependent on prior performance could better enable certain disadvantaged participants to adjust to a new platform order.
Is our research only applicable to China? Far from it. It is relevant for the planet’s 3.4 billion rural dwellers, who represent a vast pool of potential entrepreneurial talent for the digital economy. You may first think of developing countries, but most first-world economies also have rural areas that need to be revitalized. In U.S. states, such as Ohio and Kentucky, young talent leaves and hardly ever comes back. Online entrepreneurship could help lure talent, but it is important to promote a rich local information environment in these communities to nurture it. These lessons are valuable for a post-pandemic world in which an increasing number of people switch between online and offline in their work and daily routines.
Koo, W. W., & Eesley, C. E. (2021). Platform governance and the rural–urban divide: Sellers’ responses to design change. Strategic Management Journal, 42(5), 941-967.
What do we know so far, and how can we learn from international business research?
Platform papers is a monthly blog covering the latest academic research on platform competition. Blogposts are written by prominent scholars based on their research included in the platformpapers reference dashboard. This month’s blog is written by Liang Chen.
Pundits often rave about platforms dominating the digital economy. But have platforms lived up to the hype in global markets? Not quite, except perhaps for a handful of big names. Even for a posterchild like Alibaba, it was struggling right at the doorstep in Southeast Asia—the acquisition of Lazada stumbled into cultural clash and saw heavy CEO turnover in only a few years. Shopee, a start-up in Singapore, ended up becoming the region’s retail powerhouse and outgrew Alibaba’s entire international e-commerce. What went wrong? Is winner-take-all a mirage on the global stage?
International network effects, or not?
The hype about platforms is based on network effects. If foreign users and complementors are drawn to a fast-growing platform, international success should be a natural extension of domestic dominance. Conversely, when network effects decline across the border, we cannot expect platforms to dominate international markets. For some industries like ride-sharing and digital coupons, users mostly consume products and services offered by local complementors. On social media, users may be only attracted to local complements due to common preferences. Despite the frenzy about network effects, it is not rare that platforms must establish the ecosystem of users and complementors in a new country from scratch.
Multinationals vs platforms
International business (IB) scholars have been studying firms’ foreign expansion for decades. The focus has always been on multinationals—why they venture abroad, where they expand, how they enter each market, and how well they perform. A consensus arises that establishing a physical operation overseas can circumvent the transaction costs associated with cross-border dealings. Equity entry, i.e., setting up a foreign subsidiary, is considered the default mode—otherwise we would know no multinationals. But platforms are a different beast. A platform has a loosely coupled relationship with its complementors without which it is simply of little use.
While multinationals mandate what a foreign subsidiary should be doing, platforms seek to get complementors on board and prompt them to create value. Managing a subsidiary across borders is hard; coordinating a loosely coupled network of complementors surely is even harder. On the one hand, platforms rely on complementors’ local knowledge and initiative in exploring new use cases and offering locally appealing products and services. To achieve this, complementors must have substantive autonomy. On the other hand, platforms need to push complementors to make platform-specific investments. Such investments help create more value (e.g., by differentiating the platform), stronger synergies with other complements in attracting users, and they can increase switching cost for complementors. Yet only when the platform exerts a certain degree of control will complementors be willing to make such platform-specific investments. The balance between autonomy and control is a tricky one and poses an ongoing challenge in an unfamiliar market.
IB scholars traditionally depict multinationals using the notion of firm-specific advantages (FSAs). FSAs mostly derive from intangible assets, and they determine internationalization strategies and outcomes. But we all know platforms are called platforms precisely because they are not defined by proprietary assets alone. In competing with rivals abroad, platforms need ecosystem-specific advantages (ESAs) instead. ESAs arise when a platform has a great number or range of complements, when the activities within the ecosystem are highly complementary as shown by Amazon’s flywheel, and when the platform firm can govern the complementors such that their interests are aligned with the platform’s goals.
However, the transfer of ESAs across borders appears challenging. The platform will likely face a chicken-and-egg problem every time it expands into a new country. When a multinational platform cannot piggyback on its existing advantages to gain a head start, the risk of being outsmarted by a local rival becomes a real concern. After all, a local competitor will know the market better and it is not constrained by a globally consistent interface or design. Moreover, platforms may find it frustrating that even one technical bottleneck in the local market could paralyze the flywheel and deprive the platform of its ESAs. Alibaba is known for its commitment to an asset-light platform model; yet it had no choice but to invest extensively in logistics across Southeast Asia, and it also deployed the domestic logistic arm Cainiao to help enhance Lazada’s delivery efficiency. Still, the sluggish adoption of digital payment across the region poses another roadblock to leveraging Alibaba’s ESAs.
Charting new courses
As platformpapers keeps adding new publications, it is also getting clearer that our understanding of platform internationalization is still in its infancy. A closer look at IB scholarship may unveil several avenues worthy of future research endeavors.
Would we benefit from new theoretical lenses for studying multinational platforms? IB has heavily utilized transaction cost thinking in characterizing multinationals. But platforms “externalize” rather than “internalize” operations. In my research, I argue that property rights theory may be the most appropriate tool. You might agree or disagree, but the time is ripe for alternative perspectives.
In IB, local adaptation is key to success. But adapting what? On a platform, complementors will mostly take care of the adaptation of product offerings. What seems intriguing is whether and how platforms effectively adapt their governance strategies to the conditions of a new market. Failures in certain foreign markets may be attributed to an inability to adapt platform governance and design, rather than simply cultural differences.
Scholars exhibit a growing appetite for linking corporate strategy with platform research. As an example, Uber continually reconfigures its portfolio in different countries; it doubled down on ridesharing in some markets while prioritizing Uber Eats (especially grocery delivery) in others. But we don’t have a clear idea why. In Hong Kong, Uber just shut down food delivery and turned its focus to Uber Taxi. How is this different from resource redeployment? Why do platform firms diversify into other businesses (often through acquisition) in a foreign market?
Regulators always play a salient role in IB, but perhaps more so for platforms. Changes in regulation can force changes in the business model and even illegalize a platform business to the extent that it must divest (e.g., Airbnb quitting Berlin in 2016). On the contrary, Grab worked closely with Thai authorities for years—despite ridesharing being illegal—and sought support from all major parties in the 2019 general election. IB research has a lot to say about non-market strategy and may be particularly informative on how platforms can establish legitimacy in a foreign territory. We know that corporate social responsibility has been an effective means, and perhaps “ecosystem social responsibility” is the next big area for researchers to explore.
Research on emerging market firms suggests that they venture abroad to escape from underdeveloped institutions at home. Platform firms, too, may be escaping from home regulations that are either too restrictive or misaligned with other countries. It is no coincidence that Chinese household names like ByteDance, Tencent, and Alibaba all embarked on hiring sprees in Singapore lately. This also raises the overlooked question of what kind of places will be favored by platform firms for establishing physical operations.
To be sure, platform scholars are paying growing attention to country differences (e.g., regulations), and it has proved exciting to work out the various implications for platform strategy. The question is: can our research keep up with the way platform businesses expand globally? I believe that the dialogue between IB and platform researchers is no longer an option but an imperative.
This blog is based on Liang’s research published in the Journal of International Business Studies and is included in the platformpapers reference dashboard.
Chen, L., Li, S., Wei, J., & Yang, Y. (2022). Externalization in the platform economy: Social platforms and institutions. Journal of International Business Studies, 1-12.
Li, J., Chen, L., Yi, J., Mao, J., & Liao, J. (2019). Ecosystem-specific advantage in international digital commerce. Journal of International Business Studies, 50(9), 1448-1463.