How government intervention can hurt the free web

Published on in , ,


Free ad-supported business models rely on the ability of a firm to monetise user data and user attention through various forms of advertising. This model is a cornerstone of the free web and begs the question: will stricter regulation on user data protection economically impact free online web-services?

This note will demonstrate that indeed any intervention that weakens the monetisation of data and attention, will impact ad-supported online web-services in their current form. It will do so using an economic argument based on the theory of public good market failures. This note relies on a framework for classifying public goods.

Section 1 will lay the key assumptions for web access over a neutral internet. Sections 2 and 3 will demonstrate that as information goods, free online web-services can be simultaneously classified as public goods and club goods, further exploring the role of private-firm provisioning. Lastly, Section 4 will answer the stated question, by explaining how government intervention has the potential to hurt the free web that consumers enjoy today.

1. Assuming web access over a neutral network

To develop my argument, I must make two key assumptions. First, I will assume that users have readily available access to the internet. This means that users have access to the hardware, infrastructure and knowledge necessary to access the internet, or more precisely, the web. Internet access itself is a club good that is excludable by access fees, but non-rivalrous in consumption (World Bank, 2016).

Second, I will assume that access to the web occurs over a neutral network. The Electronic Frontier Foundation (EFF) describes network neutrality as “the idea that [Internet Service Providers (ISPs)] should treat all data that travels over their networks fairly, without improper discrimination in favour of particular apps, sites or services" (Electronic Frontier Foundation, 2017). It is a principle that emerged as part of the debate on “open access", which studied the vertical integration of ISPs and cable operators (Wu, 2003).

These assumptions are reasonable, especially for consumers in developed countries

Today, 95% of the world’s population lives in an area that is covered by a mobile-cellular network and 84% of global population is within range of mobile-broadband (ITU, 2016). Global internet penetration has been increasing steadily over the past 10 years (World Bank, 2015), reaching 3.45 billion in 2016 (ITU, 2016). In high-income countries, 81% of the population already has ready-access to the internet (World Bank, 2016). In developing economies, internet users per 100 people has grown 27 percent a year since 2000 (World Bank, 2015). In India for example, internet user growth is accelerating achieving 40% year-on-year growth in 2016, compared to 33% in the previous year (Meeker, 2016).

Governments in the USA (Federal Communications Commission, 2015) and EU (European Commission, 2016a) have clearly stated regulation in favour of network neutrality. The issue is also important to developing countries, since a lack of neutrality introduces foundational vulnerabilities that can be used by corrupt governments to harm their citizens. This has been recognised as a key concern in South Africa for example, where the Internet Service Providers’ Association (ISPA) published a statement endorsing network neutrality as a key tenant of democracy (ISPA, 2016).

Examples of where these assumptions do not hold include China, Turkey, and Saudi Arabia, where despite high levels of internet penetration, throttling and censorship are commonplace.

Update: since writing this, the FCC “voted to start a proceeding to repeal the 2015 network neutrality rules" (Sohn, 2017). I don’t think this hurts my assumptions.

2. Free online web-services are public goods

My first point of argument states that online web-services that are provided to users at zero financial cost (i.e. for "free") can be classified as public goods. In other words, they are non-excludable and non-rivalrous under the previously established assumptions. A proof of this will follow; first however, I will briefly describe why certain online services succeed best under a free pricing model.

Free pricing models are essential for the web

Free, often ad-supported, online web-services are a staple of web and the main reason consumers use the internet (Nie & Erbring, 2002). A pricing model that incorporates a free tier is important to online businesses for several reasons; I will highlight three.

The first reason for free pricing is that consumers’ willingness to pay for online content or services is low (Dou, 2004). There are many reasons for this, which are captured by users’ perceived benefits and risks of paying for online content. Most relevant, however, is a user’s ability to get access to an equivalent service elsewhere online for free, examples include open-source community substitutes. Therefore, the distribution of free software is often used as part of a pricing strategy to capture a user base of those with low willingness to pay.

A second reason is that software products are subject to network effects – that is the idea that a product increases in value as its user base expands (Gallaugher & Wang, 2002). Free pricing is a mechanism to boost a product’s user base and in effect increase the product’s inherent value.

Finally, online web-services can be classified as digital or information goods. These goods correspond with a theoretical marginal cost of zero (Negroponte, 1995). The economics of negligible marginal costs allow software firms to use pricing strategies such as: free, subsidised, and trialling (Gallaugher & Wang, 2002). These strategies can resolve asymmetry of information failures, such as perceived product value.

Classifying free online web-services

We will adopt a framework devised in an earlier note to classify free online web-services.

Establishing non-excludability

i. Determine all possible modes of exclusion. If a user has ready access to the web over a neutral network, then the modes of exclusion are limited. In fact, through application of the assumptions, it would be impossible for an intermediary, such as an ISP or government to exclude a user from an already delivered service. However, the web-property itself may choose to restrict user access. It can do so through access-fees or by blocking an individual user – these are the most likely modes of exclusion.

ii. Determine the rational practicality of each mode of exclusion.

a. Access fees: as demonstrated earlier, free pricing models are essential for online services. Any exclusion through financial access-fees would be irrational and potentially harmful to an online product’s core value proposition.

b. Individual blocking: while this is possible, it is often impractical to fully block an individual from an available online service. This is because of the web’s open nature and the anonymity afforded to internet users. It is especially true if a service does not require user credentials (e.g. Google search).

iii. Analyse and decide. Since all modes of exclusion are rationally impractical, free online web-services can be regarded as non-excludable. When provided, it is difficult or even impossible to exclude an individual from enjoying them.

Establishing non-rivalrous consumption

i. Determine all sources of rivalry in consumption. If a user has ready access to the web over a neutral network, then the sources of rivalry are limited to the web-property itself. A source to consider is limited server resources, such as compute-time or storage space.

ii. Determine all points-of-congestion. There is a point at which servers can become flooded and are no longer able to respond to user requests. In fact, this is sometimes used by adversaries and is known as Denial of Service.

iii. Determine if any point-of-congestion is reasonably likely, or even realistic. Modern scalable cloud infrastructure enables companies to avoid reaching this point by design. Therefore, only under exceptional circumstances (e.g. cyberattack) is such a point-of-congestion reached.

iv. Analyse and decide. Since the primary point-of-congestion is considerably unlikely, free online web-services can be regarded as non-rivalrous in consumption. One individual’s enjoyment of a web-service does not detract from that of another’s.

Since free online web-services exhibit both properties of non-excludability and non-rivalry in consumption, they can be classified as public goods. However, public goods are not typically provided by private firms, since free rider problems make charging for the good difficult. The question then arises, why do private firms provide these free online web-services? To answer this, I analysed alternative sources of revenue, namely advertising. In doing so, we find that free ad-support online web-services are actually not public goods.

3. Free online web-services are not public goods

A key premise under which free online web-services are classified as public goods is that they are non-excludable. When it is preferred to distribute software at a financial price of zero, access fees are not imposed. Instead, web-services rely on alternative sources of revenue such as advertising. However, the extent of advertising’s success is closely tied to the service’s ability to accurately profile users and monetise their attention (referred to as “eyeballs" and manifested as “impressions" and “click-through rates") (Leontiadis, et al., 2012). Companies collect vast amounts of data to enable desirable levels of profiling, a lot of which is personal.

Furthermore, the core user value on online platforms and multi-sided markets is often built on the contribution of data by users (Gallaugher & Wang, 2002). For example, a user could not derive value from Google search without first contributing a search query. Additionally, the value that a user can derive from Facebook is dependent on them first exposing interests through “Likes", or building a friendship graph by contributing data that represents their personal relationships. In summary, consumers pay for free online web-services using non-financial intangibles assets (data and attention). I therefore argue that the requirement for data and attention, presents a mode of exclusion that is evidently practical. This renders free online web-services as excludable by a practical access-requirement: the contribution of data and, subsequently or simultaneously, attention in exchange for the service. No data contribution implicitly suggests exclusion from the service, just as financial currency would for a normal good.

Free, ad-supported online web-services hence fail the test for non-excludability and should be classified as club goods. They are non-rivalrous but excludable by intangible non-financial (yet monetiseable) access requirements. Therefore, any intervention that weakens the monetisation of data and attention, will introduce public good market failures and impact ad-supported online web-services in their current form.

4. Stricter regulation on user data protection can economically impact free online web-services

The ability of private firms to provide online web-services for free, is constrained by their ability to monetise user data through advertising. The large-scale collection of data required to enable this has raised privacy concerns among consumers (Goldfarb & Tucker, 2011). This has led to government intervention (e.g. the European Union’s General Data Protection Regulation) that “enhances" user data protection by constraining how data can be controlled and processed (European Commission, 2016b). This note has adopted a unique approach to understanding the effects of this increased user data protection on ad-supported web-services, by adopting an economic argument based on the theory of public good market failures.

A framework has been applied to demonstrate that under the assumptions of a neutral network, digital or information goods such as free online web-services can be classified as public goods. Consequently, it demonstrated that the very nature of free ad-supported web-services, requires the contribution of user data. This creates a practical mode of exclusion for consumers, thereby classifying these services as club goods – they are non-rivalrous but excludable by intangible non-financial (yet monetiseable) access requirements. Thus, this note demonstrated that any intervention that weakens the monetisation of data and attention, will impact ad-supported online web-services in their current form. Highlighting the point that over-policing of user privacy may lead to public good market failures and free rider problems online. The protection of user privacy online is very important and especially relevant on today’s web, however taken to the extreme, disproportionate government intervention has the potential to hurt the free web that consumers enjoy today.

For bonus points: private firms are good at provisioning online web-services

One way to address the lack of private markets for public goods is government provisioning. An example of this can be drawn from the 2008 Quaero project – a €99M attempt by the European Commission to build a publicly-funded search engine (European Commission, 2008). The project however ended unsuccessfully in 2013 (Quaero, 2014), demonstrating that global ad-supported private markets can provide superior and more sustainable web-services online.


Dou, W., 2004. Will Internet Users Pay for Online Content?. Journal of Advertising Research, 44(4), pp. 349-359.

Electronic Frontier Foundation, 2017. Net Neutrality. [Online] Available at: [Accessed 16 February 2017].

European Commission, 2008. State aid: Commission authorises aid of €99 million to France for QUAERO R&D programme. [Online] Available at: [Accessed 15 April 2017].

European Commission, 2016a. Digital Single Market: Open Internet. [Online] Available at: [Accessed 15 April 2017].

European Commission, 2016b. Regulation (EU) 2016/679 of the European Parliament and of the Council (of 27 April 2016) on the protection of natural persons with regard to the processing of personal data and on the free movement of such data, and repealing Directive 95/46/EC (General Data Protection Regulation). Official Journal of the European Union, 119(2).

Federal Communications Commission, 2015. Open Internet. [Online] Available at: [Accessed 15 April 2017].

Gallaugher, J. M. & Wang, Y.-M., 2002. Understanding network effects in software markets: evidence from web server pricing.. MIS Quarterly, December, pp. 303-327.

Goldfarb, A. & Tucker, C. E., 2011. Privacy Regulation and Online Advertising. Management Science, January, 57(1), p. 57–71.

ISPA, 2016. Net Neutrality Indicates a Functioning Democracy. [Online] Available at: [Accessed 16 February 2017].

ITU, 2016. ICT Facts and Figures 2016, s.l.: International Telecommunication Union.

Leontiadis, I., Efstratiou, C., Picone, M. & Mascolo, C., 2012. Don’t kill my ads! Balancing Privacy in an Ad-Supported Mobile Application Market. s.l., Proceedings of the Twelfth Workshop on Mobile Computing Systems & Applications. ACM.

Meeker, M., 2016. Internet Trends 2016, s.l.: Kleiner Perkins Caufield Byers.

Negroponte, N., 1995. Being Digital. New York: Alfred A. Knopf.

Nie, N. H. & Erbring, L., 2002. Internet and Society: A Preliminary Report. IT&Society, 1(1), pp. 275-283.

Quaero, 2014. December 31st, 2013: the Quaero program comes to an end. [Online] Available at: [Accessed 15 April 2017].

World Bank, 2015. Internet users (per 100 people). [Online] Available at: [Accessed 1 March 2017].

World Bank, 2016. World Development Report 2016: Digital Dividends, Washington: International Bank for Reconstruction and Development / The World Bank.

Wu, T., 2003. Network Neutrality, Broadband Discrimination. Journal on Telecommunications & High Technology Law, Volume 2, pp. 141-176.