In their analysis of the European Digital Markets Act (DMA), Fiona Scott Morton and Cristina Caffara point out that the Act is an unstructured list of prohibition and obligations for platforms that seems to be based on past court cases, which are restated in case-independent terms.[i] Without a coherent conceptual model of the platform economy, it is hard to see how the list can be generalized to other cases. Instead of summarizing past cases, the Act should have focused on business models.
I agree. But what is the structure of Big Tech business models? What kind of anti-competitive practices do they exhibit? Are these the same for all platforms? What questions should we ask to check for anti-competitive patterns?
In this blog I define the structure of business models, provide a classification of platform business models, and identify the platform practices that the DMA seems to want to regulate.
The structure of digital business models
A business model is a statement how a business creates, delivers and captures value. In the digital economy, business models are coproductions in which a network of customers, suppliers, subcontractors, partners, complementors and other stakeholders all contribute to the creation and delivery of value, and capture benefits from this in the form of revenue or some other utility.
Digital business models do not just describe how a single organization creates, delivers and captures value, but how a value network of economic entities does this. This gives us four components of digital business models:
- An organization creates value
- with a network of economic entities
- that each must generate positive revenue from this.
- Value is delivered using digital technology.
The following business model wheel expresses that each of the four components depends on the other three.[ii]
This gives us four handles to classify platform business models.
Platform value propositions
A platform is an infrastructure in a digital ecosystem on top of which others can create additional value. For example, a digital marketplace is a platform on top of which third-party sellers can offer products to consumers, and an operating system is a platform on top of which app developers can offer apps to end users.
Platforms can offer two types of value propositions, utility services and connection services.[iii] A utility service is a service on top of which developers can offer third-party services. Examples are operating systems, game platforms, IoT platforms, content management systems and web browsers.
A connection service is a service that allows two or more parties to connect. All connection platforms connect producers and consumers of content, but the way they do this is different.
- Search engines connect users with content by answering a user query. They may personalize their service, so that different users may receive different answers to the same query.
- Marketplaces offer a search function to connect buyers with sellers of products or with providers of services. And they offer a transaction service to buy the product or order the service. Advertising networks and exchanges also fall under this category. They are platforms in which publishers of inventory (empty space to put an ad) and advertisers can connect to each other.
- Communication platforms send content to receivers specified by the sender. They do this by telephone number (WhatsApp, Signal, and other messaging systems) or by another identifier (Gmail, Outlook and other mail systems).
- Social networks match content of publishers with consumers of content, using some confidential matching criterion. This may include the publishers Friends list, the consumers expected engagement, or other criteria. Content may consist of text, images, videos, or other media. Publishers may be consumers who publish their content for free, or they may be professional publishers who receive some revenue from the social network if their content is matched. Most social networks are ad-funded.
These services need different regulations, but the DMA does not distinguish them.
Platform value models
The second aspect of digital business models are value models. Value models show what the parties in a value network contribute to delivering the final value proposition to the user. I identified four characteristic value models for different kinds of platforms in a previous blog. As an illustration, here is the value model for marketplaces. Each arrow represents a value contributed to higher layers in the diagram.
In a characteristic platform pattern, the marketplace bypasses the retailers that build value on top of it, and offers services to the customers of its retailers. As Tom Goodwin already observed in 2015, platforms own the customer interface.[iv]
The DMA contains a few regulations aimed specifically at marketplaces. Marketplaces should allow their retailers to offer the same products elsewhere cheaper; they should not be a retailer competing with other retailers on the same marketplace; and they should not preference their own products in the product search function.
Operating systems have a similar pattern but now with developers who create apps for end users. But they offer a different value proposition, so they are subject to different regulations. The DMA here forbids that the operating system refuses to run apps not obtained through its own app store, forces its own payment app on all developers, and levies an unreasonably high transaction fee on in-app purchases.
Platform revenue models
The third aspect of business models is the revenue model of the participants in the value network. A revenue model shows how each of the participants generates revenue from their participation in the network. Here is the revenue model of Amazon Marketplace, using the e3value notation.[v]
The upper half of the model says that buyers have a need, represented by the bull’s eye, which they can satisfy by buying a product from a seller. The lower half says that when this happens, the Marketplace facilitates the transaction, collects data about the buyer and seller, and is paid for this by the seller. The buyer adds value to the Marketplace simply by being present.
Models like this force us to make the value exchanges among network participants explicit. They can be used by network participants to assess the revenue generated by their participation in the value network, and by regulators to assess fairness of competition. In the case of marketplaces, regulators may wish to impose some boundary conditions on the revenue model, such as on the size of the transaction fee and on transparency of data collection.
Two blogs ago I provided a catalog of platform revenue models: donations (e.g. Signal), subscription (e.g. premium services of LinkedIn), and usage-based models (e.g. a transaction fee for every sale on a marketplace). Ad-funded platforms ask a fee for every impression of an advertisement, which is a special case of usage-based funding.
For each of these revenue models, regulators may want to issue some rules. For example, the source of donations must be transparent, transaction fees must not be outrageously high, and the performance of ads must be independently verified. The DMA contains some rules for ad-funded revenue models but misses the most important pattern, namely that some ad-funded platforms are buyers, sellers, and marketplace operators at the same time.
Platform delivery models
The delivery technology of digital platforms is, well, digital. Some elderly lawmakers have started to realize that computers are more than automated typewriters. But the following feature of the digital economy still needs to sink in:
Digital service delivery = digital data collection.
The details about what data is collected and how this benefits the platform’s revenue model may be hidden in technology. Not all participants in the value network may understand the network’s delivery model, hence fail to understand the platform’s business model, and therefore misunderstand the distribution of benefits in the platform’s value network.
For example, Meta, the company formerly known as Facebook, offers APIs through which developers can access Facebook’s social graph and share content with it.[vi] Facebook Connect allows end-users to log into another site, outside the Facebook app, using their Facebook ID.[vii] Social plugins allow a publisher site to include Facebook buttons (like, share) on their site. The Open Graph Protocol allows sites to become part of the social graph.[viii] All of these delivery technologies greatly enhance the capability of Meta to collect data about users inside and outside Facebook. Businesses who use these technologies may not all be aware of the huge revenue boost they give Meta.
The story of Facebook Instant Articles, which extracted value from news publishers without giving anything back, is another illustration of unequal distribution of benefits hidden in delivery technology.[ix] The complexity of ad-funded business models, which extracts a lot of value from news publishers too, is another illustration.
Anticompetitive practices in the DMA
If we follow the classification of platforms according to business models given above, the DMA becomes more understandable, and it becomes clearer how the DMA applies to new cases. Here is what it says when rephrased in terms of our business model structure:
- Marketplaces should allow their retailers to offer the same products elsewhere cheaper; they should not be a retailer competing with other retailers on the same marketplace; and they should not preference their own products in the product search function.
- Operating system providers that also develop apps for their operating system, should not make their own apps non-removable and default, they should run competing apps, and allow easy switching to these other apps. They should run apps not downloaded through their own app store, allow non-native payment systems and charge reasonable transaction fees.
- Ad-funded platforms should not run their own ad exchanges, not force companies to use the platform’s reader identification system and be transparent about pricing and performance of ads.
There are some more prohibitions not related to the platform business model, such as the prohibition on bundling and on restricting business users from raising issues with relevant public authorities.
It remains to be seen how the DMA works out on legal practice. However, there is hope on the other side of the ocean: The AMERICA act (Advertising Middlemen Endangering Rigorous Internet Competition Accountability) proposed by a bipartisan committee led by senator Mike Lee is an shining example of the needed focus on business models. It prohibits the conflict of interest in ad-funded platform business models, and nothing else. The ad-funded platforms run by Google and Meta are publishers offering inventory, represent other publishers who offer inventory, also represent advertisers who buy inventory, and run a marketplace where publishers sell ad inventory to advertisers. The EU should learn from this example and rephrase the DMA in terms of digital business models.
[i] Fiona Scott Morton & Cristina Caffarra. “The European Commission Digital markets Act: A Translation.” VoxEU/CEPR. 5th January 2021. https://cepr.org/voxeu/columns/european-commission-digital-markets-act-translation
[ii] Roel Wieringa & Jaap Gordijn. Digital Business Ecosystems. How to Create, Deliver and Capture Value in Business Networks. TVE Press, to appear, 2023.
[iii] Ibid. Chapter 3.
[iv] Tom Goodwin, “The battle is for the customer interface,” TechCrunch, 4th March 2015, https://techcrunch.com/2015/03/03/in-the-age-of-disintermediation-the-battle-is-all-for-the-customer-interface/.
[v] J.Gordijn & R.J. Wieringa.E3value User Guide. The Value Engineers, 2021.
[vi] Facebook. “Facebook unveils platform for developers of social applications,” May 2007. https://about.fb.com/news/2007/05/facebook-unveils-platform-for-developers-of-social-applications/.
[vii] Dave Morin. “Announcing Facebook Connect.” May 2008, https://developers.facebook.com/blog/post/2008/05/09/announcing-facebook-connect/.
[ix] Roel Wieringa & Jaap Gordijn. Digital Business Ecosystems.