David vs Goliath or Standing on the Shoulders of Giants?

– A comment on The Digital Markets Act (DMA) with respect to SMEs and startups

–
Joakim Wernberg, PhD, Research Director of Digitalisation and Tech Policy, Swedish Entrepreneurship Forum

Senior Lecturer at the institution for Technology and Society, Lund University

>> Download PDF

–

The European Commission is in the process of adopting new regulation for the digitalised economy, of which the Digital Markets Act (DMA) sets out the fundamental rules for digital platforms and platform companies. While the purpose the DMA is to advance a digital single market in the EU and a level playing field for those acting on it in the future, the proposal itself, as well as the policy debate surrounding it, has centred almost exclusively on the regulation of large tech companies, such as Google, Amazon, Facebook, Apple and Microsoft. There appears to be a considerable interest in casting the DMA as a story about David vs. Goliath, where regulators need to step in to defend smaller firms against the tech giants. This is not necessarily the case, and it would be more in the interest of the market to think of it instead as story of the potential that lies in standing on the shoulders of giants.

By regulating the largest platforms, legislators hope to enable smaller firms to grow and promote new market entry. However, the DMA fails to account for the fact that multi-sided platform economies and digital services differ in significant ways from traditional businesses.[1] Consequentially, the proposal instead comes with the considerable risk of unintended negative impact on small- and medium-sized enterprises (SMEs) as well as on the next generation of tech-driven startups and platform companies.

This essay covers five aspects of the DMA that may impact SMEs and tech startups negatively and proposes corresponding measures to address some of these issues in the ongoing negotiations to improve the EU’s future digital markets:

1. Enable, rather than inhibiting, the building of ecosystems of platform services through interlinking and data sharing. This lowers the barriers to SMEs using platform services and promotes innovation.
2. Add the requirement that a platform has to hold a dominant position in its market to be designated a gatekeeper. This ties the DMA to existing antitrust regulation and promotes a case-by-case approach that takes platforms’ heterogeneity and specific multi-sided characteristics into account.
3. Shift the criteria for reporting acquisitions from turnover to the price of the acquisition. That way, the regulation is used to prevent anticompetitive behaviour while also safeguarding a market for entrepreneurship and innovation.
4. Provide clear definitions and delineation between personal data and structured data to simplify compliance and safeguard privacy as well as business secrets.
5. Limit the mandate provided to the EU Commission based on delegated acts in favour of clearly defined and predictable long-term rules for the market.

The rest of the essay is divided into five parts addressing the definition of the digital sector, the role of size in competition, startup acquisitions, the value of data, and long-term rules for the market. This is followed by some concluding remarks.

1. Beyond the digital sector

The DMA is being put forward as a sector-specific piece of legislation limited to what is described as the digital sector, but this paints a misleading picture of the state of digitalisation and the role of digital platforms in the economy. Just as there is no longer a meaningful distinction between being online and offline, the economic impact of digitalisation stretches across sectoral boundaries and permeates the entire economy. Consequently, regulation of the digital market will have far-reaching effects across the economy, including technology-intensive startups as well as small and medium-sized enterprises that are less digitally advanced.

The business of multi-sided platform economies is essentially matchmaking between groups on different sides of the platform.[2] While increases in digital connectivity has led to an increase in the number of potential matches between supply and demand, multi-sided platforms play a key role in enabling the realisation of this potential through matchmaking and lower transaction costs. Put differently, without digital platforms we cannot leverage the benefits of digital markets.

There are digital platforms operating across a wide and growing variety of sectors both within and across national borders. While firms like Facebook, Google and Amazon provide unprecedented market access through increased geographical reach in real time, other platform companies like Uber, Bolt or Delivery Hero increase market access by matching supply of and demand for services that are highly localised and limited in time. A platform like Instagram can be used to promote a new line of oxford shirts to customers across Europe or to offer a discount price on fresh pastries at a local bakery the hour before closing for the day.

Behind the remarkable ad revenues and profits of many of the largest platform companies lies a second tier of economic analysis that seldom makes the headlines, namely the revenues and profits of the many companies that use these platforms to advertise and to do business. This is especially true for small and medium-sized firms, a majority of whom are digital laggards across Europe.

In 2020 we conducted a survey among 5 385 small and medium-sized firms in Sweden to gauge how they were impacted by the pandemic and what role digital transformation might play in their post-pandemic recovery. The results show that while most SMEs have narrow margins for investing in digital transformation, the top three areas, by more than ten percentage points, prioritised by most firms for building back better after the pandemic are social media (36%), e-commerce (23%), and cloud services (23%) – all of them provided by large platform companies.[3] OECD studies indicate that the productivity benefits leveraged from digital platforms are disproportionately larger for smaller firms, leaving a considerable potential to be realised as more SMEs take this step in their digital transformation.[4]

Large digital platforms allow SMEs to buy into benefits of digital technologies “off the shelf”, otherwise only attainable through own development by the largest companies. Consequently, digitally lagging SMEs are likely to be directly affected by the DMA, despite it being aimed at a vaguely defined digital sector.

The DMA recognises that large digital platforms play a vital role in providing market access for many business users and identifies the platform companies as gatekeepers. At the same time, the proposal fails to reflect how the same business users will be affected by the proposed regulation. Restrictions both in the original proposal and in subsequent amendments on ecosystems of services (Article 5 and 6) – the interlinking between services provided by the same platform company, sharing of data between services and some instances of self-preferencing – would de facto lower usability and raise barriers to SMEs using these ecosystems of services today. Such restrictions also limit the scope of innovation to offer new services in the future. This will also likely result in a corresponding decrease in consumer welfare.

The practices in article 5 and 6 do not correspond to standard “black lists” items that have an exclusively negative impact on competition and thus they require a more balanced treatment. It is worth remembering that the platforms being identified as gatekeepers have reached this position because they have expanded market access for others, especially for SMEs, in unprecedented ways, and not because they have limited it, as the DMA legislation would suggest.

2. Platform size and competition

The DMA assumes that the size and market concentration of the largest platform companies is evidence of a lack of competition and an indication of anti-competitive effects. However, considering that multi-sided platforms differ from other business models, it requires some scrutiny.

Platforms grow by leveraging network effects that increase the matching quality of the platforms service, for example the number of friends on a social networking platform or the number of potential dates on a dating platform. On the other hand, there are also negative network effects that counteract platform growth through crowding and noise that decrease the platform’s matching quality. Digital platforms will grow as large as they can by balancing positive network effects against negative ones.[5]

This holds three important implications for platform competition. First, it is hard to grow new platforms to the critical level where positive network effects can be leveraged to scale it up.[6] Therefore, many new platforms start out by targeting a niche audience, from which they can later expand. Second, as platforms become very large, it becomes increasingly hard to balance positive network effects against negative ones.[7] Large platforms use newsfeed and recommendation algorithms or try to get users to interact in smaller groups and subnetworks in order to maintain and grow their matching quality in the presence of increased crowding and noise. Consequently, multi-sided platform markets tend towards skewed market distributions with a small number of very large platforms followed by a long tail of small platforms. More importantly, this is beneficial to competition on the platform and consumer welfare.

If an additional bookstore opens in a middle-sized town with only one bookshop, the economic theory of competition indicates that competition between the two will drive prices on books down to the benefit of consumers. If, in the digital world, there is one platform that connects bookstores to consumers and that market is instead split between two platforms, the competition on each platform goes down and transaction costs go up. To achieve the same level of competition and matching quality, both consumers and bookshops would need to be on both platforms. Thus, SMEs as well as consumers benefit from the skewed market distribution between platforms.

While this is a winner-takes-most market, market distribution is not the same as market dynamics. The mere size of the largest digital platforms creates an end-of-history illusion, making it hard to believe that they can be outcompeted, but they can.[8] Competition in digital platform markets is likely to grow from a niche position in the market and topple the largest incumbents by introducing radical innovations that they are unable to imitate and integrate into their own platforms. This is how Google outcompeted AltaVista and Facebook ousted MySpace.

Apart from their size, a closer look at large digital platforms may reveal them to be more different than they first appear, not least because of how they have adapted to the markets they operate in. Reigning them in under the same set of regulations implies a delicate trade-off between, on the one hand, the range of platforms or markets being targeted by the regulation and, on the other hand, the scope of regulatory intervention. The DMA proposal accounts for this heterogeneity in the criteria used to identify gatekeepers that fall under the obligations of the law. A mix of quantitative metrics and qualitative judgement is suggested in order to be able to include platforms that may otherwise fall outside of the scope of the regulation. Yet, the proposal does not reflect the corresponding risk that a set of metrics used across different markets will have a skewed impact on heterogeneous platforms.

One way to address this issue would be to add the requirement that a platform has to hold a dominant position in its market to be designated a gatekeeper. This would take differences between markets and the heterogeneity of platforms into account, while also strengthening the connection between the DMA and existing antitrust regulation. This enables a case-by-case approach from antitrust authorities in order to fully take differences between heterogeneous platform businesses into account and improve the overall understanding of their multi-sided nature. It is also important to underscore that existing antitrust regulation applies to the large platform companies, it works, and there are indications that its applications in this area are far from exhausted.[9]

3. Tech startups and acquisitions

Large digital platform companies play a key role in lowering the barriers to entry for new tech startups, especially for software-based ventures and digital services.[10] With cloud services and Software as a service (SaaS), new firms can leverage resources corresponding to unattainable investments in physical capital at a low and adjustable cost. Using social media and app stores, they have access to marketing and sales channels with potentially global reach. At the same time, there is increasing worries that large tech companies are inhibiting innovation, entrepreneurship, and competition by buying up small startups.

This debate on “killer acquisitions” is lop-sided and emphasises the narrative that acquired startups would otherwise have grown into new, independent large firms and potential competitors to the incumbent platform companies. On the other hand, large tech firms’ acquisitions complement venture capital and other sources of startup funding. [11] They also provide a viable exit strategy for entrepreneurs that do not wish to grow one single venture but go on to invest in new ideas or pursue other activities. Thus, large firms acquiring startups contribute to broadening the variety of entrepreneurial effort that is being put into action in the economy.

Empirical research on killer acquisitions – i.e. big large companies buying up small innovative firms and scrapping their innovation – comes largely from the pharmaceutical and biotech industry and shows some small indications of anticompetitive conduct but not on a systematic scale. More importantly, the differences in barriers to entry makes it an even less viable strategy in digital platform markets where it is much easier to start a new venture.[12] More likely, large tech companies are acquiring small innovative firms as a means to remain innovative themselves.[13] In doing so, they also create a market for ideas and entrepreneurship, as evidenced by the wide variety of incubators, accelerators, VC funds, communities and new entrants in software-based ventures and digital services. The risk of anticompetitive behaviour must be balanced against the gains of a widened scope for entrepreneurial ventures.

Against this backdrop, the proposal for amended controls of mergers and acquisitions in the DMA needs to consider the potentially negative effects on innovation and entrepreneurship brought on by increases in the number of reviews and lengthy review processes. An alternative approach worth exploring would be to shift the criteria for reporting acquisitions from turnover to the price of the acquisition. This reflects the value that the acquiring firm puts on the acquired firm, as well as the entrepreneur’s valuation of the foregone profits from staying to compete in the market. That way, the regulation is used to prevent anticompetitive behaviour while also safeguarding a market for entrepreneurship and innovation.

4. The volume, structure and value of data.

A central theme in the DMA is that data is a crucial resource for innovation that is unevenly concentrated among the largest platform companies. Therefore, gatekeeper firms should be obliged to release throws of data that can be used by new entrants and smaller firms to innovate. There are three principal problems with this approach:

First, the value of data does not lie in its volume, which is cheap to scale, but in its structuring, for which the costs do not scale as favourably with size.[14] This implies that different firms may extract different value from the same data set and that it may require more resources to extract certain types of value from very large data sets. Thus, data-driven innovation may not become more evenly distributed just because data is made a public resource. Second, the work that goes into refining data into structured data is at the core of every data-driven enterprise and constitutes part of its business secrets. Third, a large part of the data that gatekeepers govern is personal data which is restricted by the agreement between the platform and its users as well as by legislation such as GDPR.

The proposals in the DMA are pitting both privacy and business secrets against the hope that the published data will generate future innovation. This may incur serious harm to incumbent platform companies (gatekeepers), but it also constitutes a serious risk for emerging platform companies who, upon becoming gatekeepers, would be required to amend the agreements with their users, concerning data as well as the risk of having to publish business secrets. Between personal data provided by the users and structured data that is part of the firm’s intellectual property, there is no clear definition of what data gatekeepers should be expected to release. Furthermore, the DMA should take into account that the openly published data may also be used for nefarious reasons. Even aggregated and anonymised data sets that are openly published can be interlinked with other data sets in ways that may pose a risk to privacy.[15]

If the purpose is to enable startups and smaller firms to engage in more data-driven innovation, rather than to focus on the volume of data held by large tech companies, the DMA should narrow in on the definitions and delineation between personal data and structured data so that compliance with existing regulation becomes more straight-forward. Smaller companies cannot afford the investment in compliance that larger firms can, and therefore this creates a hurdle of uncertainty for them. A clearer distinction between personal data and intellectual property also makes it more straight-forward to require that users always can export their personal data and move it to other platforms, should they wish to do so. This would not only add to user privacy and allow users to favour new services by porting their personal data, but also help protect and govern business secrets in a way that promotes further investment in data-driven innovation.

5. Long-term rules of the market

The DMA relies on ex ante regulation, imprecise or flexible criteria for designating gatekeepers, and a significant shift in power to the European commission through delegated acts in order to allow for changes of gatekeeper criteria and obligations. Taken together, this implies that the DMA favours regulatory mandate at the detriment of long-term predictable rules and transparency in the market. The DMA adds structural uncertainty to platform businesses because the criteria of inclusion and the obligations of, as well as and sanctions for, gatekeepers are suggested to be subject to additional individual assessment and changes mandated by delegated acts. Consequently, not only will the directive contribute to a regulatory moat, but the cost of crossing that moat will also be hard to estimate. This uncertainty affects both tech startups and less digitally savvy SMEs.

SMEs using digital platforms to do business are part of a changing business structure, characterised by increased and deepened interdependencies between firms in the economy. This calls for updated legislation, but the DMA recognises only half of this relationship – it acknowledges that smaller firms may be dependent on large platforms but does not take into account how the proposed regulation aimed at platform companies may lead to unintended negative effects for smaller firms relying on these platforms. Similarly, increased uncertainty for platforms companies translates into increased uncertainty for the firms using the platforms. At worst, this may further impede the digital transformation of European SMEs.

Concerning tech startups, there is an inherent conflict in the DMA. While it is designed to promote new (European) tech champions to challenge companies like Google, Facebook and Amazon, the criteria used to identify gatekeepers and impose stricter regulation on them correspond to a large degree with the characteristics of so-called unicorn startups. This implies that the European champions that many policymakers are hoping for will be negatively impacted by the laws enacted to pave the way for them.[16]

One approach to lowering the level of uncertainty associated with the DMA would be to limit the scope of the mandate delegated to the European Commission through delegated acts. This decreases the flexibility of the regulation but contributes to more transparent and reliable rules.

Furthermore, the DMA primarily targets US tech companies, putting strain on EU-US trade relations. If enacted, it is likely to lead US to claim that EU is engaging in discrimination and protectionism in violation of WTO agreements.[17] Such tensions will affect the entire economy, but SMEs or tech startups hoping to access the US market will be especially vulnerable.

Concluding remarks

The purpose of the DMA is two-fold. First, to prevent further fragmentation in Europe´s digital markets by foregoing national legislation aimed at platform companies. Inc ombination with further harmonisation of regulatory frameworks across EU member states, this is an important policy priority that will benefit both platform companies and the economy as a whole.

The second purpose of the DMA is to promote fair competition and contestable markets for platform companies and the services they provide. Rather than being an end in itself, competition is usually thought of as a means to achieve new market dynamics, new market entry, innovation, and – by extension – consumer welfare. More specifically, the proposal rests on the assumption that the mere size and revenue of the largest platform companies is evidence of anticompetitive behaviour or anticompetitive effects. As I have shown in this essay, multi-sided platform economies differ in important respects from traditional business models, meaning both size and market concentration may in fact be beneficial to business users and consumers as well as to future platform companies.

The DMA should not be cast as a story of David vs. Goliath where regulators need the come to the aid of smaller competitors, but about the innovation and entrepreneurship that can come from standing on the shoulders of giants. For this to happen, the proposal must take the impact on SMEs and startups into account. Competition between platform companies will have its David and Goliath moments with skewed market distributions and scaling up newcomers, but that does not mean regulators have to intervene. The story behind the metaphor should help us to remember not to overestimate size or to underestimate innovation. This essay has presented five suggestions for measures towards this goal that deserve further elaboration in the forthcoming negotiations.

–

[1] Wernberg, J. (2021). Innovation, Competition and Digital Platform Paradoxes. Policy Papers on Technology, Ecnomics and Structural Change. Swedish Entrepreneurship Forum

[2] Evans, D. S., & Schmalensee, R. (2016). Matchmakers: The new economics of multisided platforms. Harvard Business Review Press.

[3] Wernberg, J. (2020). Små och medelstora företags digital omställning efter pandemin. Swedish Entreprenurship Forum.

[4] OECD (2021). The Digital Transaformation of SMEs. OECD Studies on SMEs and Entrepreneurship.

[5] Evans, D. S., & Schmalensee, R. (2013). The antitrust analysis of multi-sided platform businesses (No. w18783). National Bureau of Economic Research.

[6] Rochet, J. C., & Tirole, J. (2003). Platform competition in two-sided markets. Journal of the european economic association, 1(4), 990-1029.

[7] Wernberg, J. (2019). I den svarta lådan: Plattformsekonomier och digitalisering. In Plattformssamhället: Den digitala utvecklingens politik, innovation och reglering (pp. 22-60). Fores.

[8] Evans, D. S., & Schmalensee, R. (2016). Matchmakers: The new economics of multisided platforms. Harvard Business Review Press.

[9] De Streel, A., & Larouche, P. (2015). Disruptive innovation and competition policy enforcement. OECD DAF/COMP/GF(2015)7

[10] https://www.economist.com/special-report/2014/01/16/a-cambrian-moment

[11] Petit, N., & Teece, D. J. (2021). Innovating big tech firms and competition policy: favoring dynamic over static competition. Industrial and Corporate Change.

[12] Madl, A. C. (2020). Killing Innovation?: Antitrust Implications of Killer Acquisitions. JREG Bulletin, 38, 28.

[13] Akcigit, U., & Kerr, W. R. (2018). Growth through heterogeneous innovations. Journal of Political Economy, 126(4), 1374-1443.

[14] See for example: https://fs.blog/the-big-errors-of-big-data/ and https://www.wired.com/2013/02/big-data-means-big-errors-people/

[15] https://www.wired.com/2007/12/why-anonymous-data-sometimes-isnt/

[16] https://www.euractiv.com/section/digital/news/macron-wants-europe-to-have-10-tech-giants-worth-e100-billion-by-2030/

[17] https://www.csis.org/analysis/implications-digital-markets-act-transatlantic-cooperation