Skip to main content
Algorithms and competition law text on black and white lady's face

Algorithms and Competition Law

Digital Platforms & Algorithms: Beware the Competition Risks

Algorithms and competition law go hand in hand, yet few people realise the legal dangers involved. Given that many tech businesses aim for market disrupting, super fast growth, they are often exactly the sorts of businesses that the CMA shows an interest in. It is also worth noting that, once the CMA has begun an investigation about one issue, they go through every business record with a fine tooth comb and they will look for, and find, evidence of all and any breaches of competition law. So it is vital for fast growing, ambitious tech businesses to have a clear understanding of all the aspects of competition law.

For our American friends, competition law in your country is know as anti-trust law. 

Digital Advancement and Competition Law

The defining trend of the last decade has been the rise of the digital platform. This business model provides slick, low-touch services to customers through mobile and web applications. The logistics of fulfilling the service are typically powered by highly-automated software services, often underpinned by AI. As we have written elsewhere, these platforms have had a radical and disruptive impact on industry after industry as customers have left established businesses in favour of slicker, better experiences.

However, the Competition & Markets Authority (CMA) has now fired a warning shot across the bows of highly-digitised businesses. A new report (Algorithms: how they can reduce competition and harm consumers (publishing.service.gov.uk), published by the CMA in January, explains that algorithms have the potential to do tangible harm to consumers and, in the worst case scenarios, to directly impede competition. 

As digital innovation progresses rapidly, it outpaces the progress of the legal system. The legal system is therefore relegated to playing catchup: it simply cannot legislate quickly enough to keep up with technological advancement. Significant legal development is therefore left to case law, in the form of test cases. Test cases are, by their nature, on the cutting edge of legal development. They are therefore incredibly expensive to defend, highly uncertain and a huge drain on company and personal resources. Test cases are selected by CMA. In order to preserve their resources they choose the cases they have the best chance of winning. A good working knowledge of competition law will help you avoid being the next test case.

algorithms and competition law advancement needs test cases

In this article, we explore some common digital practices and the risks that these pose with risk to competition law, including algorithms and competition law.

What Is Competition Law & Why Should I Care?

Entrepreneurs could be forgiven for assuming that competition law only applies to bluechips and multinationals. We tend to think, for example, of Rupert Murdoch’s repeated attempts to acquire Sky News which were blocked on the grounds they had the potential to impede media plurality (Rupert Murdoch’s Sky bid is not in public interest, says regulator | Sky | The Guardian). However, as explained here (Competition law - the basics (pinsentmasons.com)), competition law applies to any business regardless of their legal status, size or sector. If this article raises any concerns that you may have you should, therefore, seek professional advice.

In general, though, there are some key principles to consider when designing and implementing your business services and practices. Embedding these into your organisation will ensure you remain competitive and on the right side of the law.

Algorithms and Competition Law: The Underlying Principle

As the UK Government explains (Cheating or Competing? Its your business to know the difference. (campaign.gov.uk), the most important takeaway is that competition is fundamentally good for customers. Therefore, building and maintaining a free and competitive market has been a global priority for governments for at least the last thirty years. It is notable that, in the UK, both the incoming Labour government of 1997 and the incoming Conservative government of 2010 implemented significant reforms to competition law in their first terms in office.

The principle to remember is that effective competition between businesses helps to keep prices fair and quality high for customers. As we will discuss, there are a range of practices which businesses can implement to shut competitors out of their market and therefore increase their own access to customers. These practices have the effect of reducing customer choice and therefore remove the need for businesses to compete on price or quality. Competition law combats this effect by making certain practices illegal, with significant fines and potentially prison sentences for those who break the law.

Anti-Competitive Practices

Given the potential commercial risks that individuals can face for breaking competition law, entrepreneurs and business leaders need to have a good understanding of the kinds of practices that may be considered anti-competitive. For those who would like a deeper dive into what the law says, there is a good overview here (https://hjsolicitors.co.uk/article/competition-law-an-overview/), but there are essentially two key practices which all businesses need to be aware of, and avoid.

Business Cartels

The first of these practices can arise where businesses reach an agreement with one another which has the effect of reducing competition. This may be an agreement to:

  • Fix Prices by agreeing to maintain prices above a stated minimum level, or by agreeing to co-ordinate price increases with one another;
  • Share Markets by agreeing who will operate in certain defined territories;
  • Rig Bids by co-ordinating amongst themselves when bidding for business to determine who will win the bid and at what price.

Organisations which engage in such practices are referred to as “cartels” and the individuals involved in these agreements can face significant fines and prison sentences.

There are some exceptions to this rule (for example, where companies have less than a 30% market share, or where reducing competition can have wider economic or social benefits). However, the burden of proof lies on the person seeking to rely on the exemption and so you need to be sure that any exemption applies before adopting the practice.

It is also worth noting that an “agreement” doesn’t have to be in writing, or even formal, to breach competition law. For example, in the case of the concrete cartel (Concrete companies: construction cartel - Case study - GOV.UK (www.gov.uk)) several businesses in the concrete sector met to carve up markets and fix prices. No formal agreement was reached but discussions of their intentions was sufficient to lead to a prosecution.

Worse still, an agreement need not be directly made between two undertakings to be considered an anti-competitive agreement. In the case of the replica football kits, the Competition Appeal Tribunal established () the important principle that an anti-competitive agreement may arise merely from one business signalling to another, via an intermediary, its future intentions with respect to price or market positioning. In this scenario, it would be very difficult for the party receiving that information not to act on it (for example, by adjusting its prices accordingly) and therefore co-ordination may be considered to have occurred.

Abuse of Dominance

The second practice which may run the risk of being considered anti-competitive can arise where a single organisation has a dominant position within its market. There is an immediate complexity here for UK businesses because, as this article explains (Abuse of a Dominant Position – InBrief.co.uk), determining whether a business has a “dominant” position depends on a variety of complex factors. The general rule of thumb is whether the company has a 50% market share or not, although companies with 40% share of the market have also found to be dominant. It is also not always straightforward to determine what segmentation may be a “market” for the purposes of assessing market share.

Having a dominant position in a market is not in and of itself a risk. The risk arises where an organisation abuses its dominant position to restrict trade, thereby giving themselves an unfair advantage. Dominant market players may achieve this in several ways, for example by:

  • Applying different trading conditions to equivalent customers in order to place some customers at a disadvantage;
  • Imposing unfair price or trading conditions on other businesses;
  • Erecting excessive barriers to entry which prevent new businesses competing in the market;
  • Limiting production or innovation within the market to the disadvantage of consumers.

use competition law to your advantage

Fair Competition Is Good for Businesses Too

If this is news to you (and, let’s face it, it’s not exactly common knowledge), then it’s worth remembering this: successful businesses grow by delivering better products and services to your customers at competitive prices, not by inhibiting the ability of their competitors to operate.

In our experience, competition law is not just good for consumers, it is also great for SMEs and start-ups too since the law lowers barriers to entering markets and requires dominant players to treat you fairly.

To give one example, we were hired to provide growth services to an educational start-up based in the UK. The business was licensed within the UK to 'organise and present' certain courses and certification within the UK and Ireland. Initially the courses were hosted at physical venues but, to increase our clients’ scalability, we supported them in moving their courses online.

One side effect of this was that our client received a lot of passive sign-ups from outside of the UK, accelerating the growth of the organisation. The IP distributor (who had granted the license to our client) challenged them. They demanded that they stop accepting sales from outside their territory. Our client was able to defend themselves on two grounds: first, that the contract only restricted where courses could be 'organised and presented' and secondly, that if the contract were interpreted as restricting sales, the effect would be anti-competitive which would be unlawful. The contract also contained a clause that stated that any part of the contract deemed unlawful would be severable and treated as though it had never existed.  This is one of many examples where competition law can be used by young, growing businesses to ensure that their growth is not restricted. 

Algorithms and Competition Law: Are Digital Platforms Anti-Competitive?

are your algorithms unlawful yellow text on black and white background

Competition law, then, is a powerful tool that global governments are using to stabilise market conditions. Given the disruption that digital platforms have driven across almost every industry in the last decade, it was inevitable that competition regulators would, sooner or later, want to take a critical look at the various digital business models.

Of course, the use of algorithms to automate business services or drive operational efficiency is not inherently anti-competitive. However, the fundamental structure of many of the most successful digital platforms is that they provide consolidated marketplaces which connect customers with pools of suppliers, where the suppliers are generally not related to one another. Amazon, for example, enables businesses to sell their products through the Amazon Marketplace. AirBnb provides a platform for independent hospitality providers to promote their businesses. Food vendors can sell through Deliveroo. And so on.

The challenge that these models create, from the competition perspective, is that they transfer power to regulate markets away from the competition regulator and into the hands of the platform vendor. The platform typically decides which products and vendors the consumer sees. The platform also has access to unprecedented volumes of data about the pricing models and selling intentions of large numbers of independent businesses.

Furthermore, digital platforms often use machine learning (ML) to determine which products to display and, in some cases, the price of those products. ML, however, has a tendency over time to converge towards the most popular products and so to discriminate against smaller vendors.

How the platform provider uses this information to affect the products and pricing that the consumer sees can have a significant impact on consumer choice. Since the algorithms and source code used by digital platforms to determine which choices to offer consumers are the platform provider’s intellectual property, the process is generally far from transparent.

Digital business models can easily give rise to competition risks in several ways:

Self-Preference May Be Abuse of Dominance

One issue that arises, particularly where the platform provider uses its own platform to sell its own products or services, is that of self-preference. For example, Google was fined 2.4bn EUR in 2017 for abusing its dominance with respect to its shopping service. Previously, when Google users searched for products to buy on the search engine, Google’s own shopping service was displayed at the top of the search results. The European Commission concluded Google’s search engine had breached its dominant position by unfairly promoting its shopping service to the disadvantage of other price comparison sites.

Similarly, the CMA report raises concerns over Amazon’s buy box which displays the default seller for a given product in search results. The CMA notes that there is a lack of transparency as to how this seller is chosen, and that it is often Amazon itself.

The key issue here is that where a platform provider is also a vendor on its own platform, a conflict of interest arises. On the one hand, the platform provider has an obligation to create a fair and equitable marketplace which works to the benefit of consumers. On the other hand, the platform provider is clearly motivated to promote its own products and services to consumers. However, where the latter concern becomes the dominant driver of decision-making, a risk will likely arise that the platform provider will abuse its position.

Price Automation Could Be Price Collusion

A second issue potentially arises with respect to pricing. Broadly, there are two different approaches that digital aggregators take to pricing: to aggregate prices from vendors and charge commission, or to set the prices that consumers pay. These practices may lead to price collusion between competitors in at least two ways.

Firstly, where price is considered as a factor when determining the rankings of search results, there is a risk that the platform will artificially inflate prices. As an example, the CMA notes the allegation that some hotel booking platforms impose price parity restrictions on vendors by demoting operators in their rankings if they offered their services cheaper on competitors’ platforms. If true, this could amount to deliberate price collusion between booking platforms because platforms imposing such constraints would effectively be agreeing with one another not to undercut one another.

Secondly, the price aggregation model creates a risk of hub-and-spoke collusion. Many platforms offer tools to vendors which enable them to determine optimal prices for their products when selling on the platform. Furthermore, some platforms provide an option for vendors to allow the platform to set their prices automatically. The CMA notes that this may amount to criminal collusion because the platform provider is essentially facilitating price co-ordination between its vendors. The key question appears to be whether the recommendations offered to each vendor are calculated independently or are simply based on the pricing offered by other suppliers.

Integration Could Be A Barrier to Entry

A further issue that arises is that some platforms offer services are useful because of their ability to automate parts of the service that their customers provide. For example, communication providers offering services such as bulk messaging or SMS capabilities are most effective as part of a wider service. In such cases, the customer of the platform provider will want to integrate their own technology with the platform that they have purchased.

This creates an opportunity for the platform provider to create barriers to entry by, for example, making their APIs excessively complex to consume. Where this is the case, there is once again a risk that the platform provider has abused their dominance by making it excessively costly for smaller customers to use their products. In some cases, this may have the effect of placing smaller customers at a disadvantage since it has the potential to impact their ability to offer the sale quality of service as their larger competitors.

Conclusion: Design Principles for Fair Algorithms

It is clear from the CMA’s report that the transparency and fair decision criteria are the key themes here. Remember, the goal of competition law is to protect the consumer’s right of choice, quality, and fair prices. As such, there are several design principles that digital businesses should adopt when designing and implementing their platforms.

Firstly, where a digital business provides a gateway for multiple vendors to sell their products, the platform provider should seek to implement Chinese walls between vendors to avoid inadvertently enabling practices which may be deemed as collusion. The key principle here is not to use information provided by one vendor in order to make recommendations about positioning, pricing or marketing to another vendor. This would require, for example, that the platform provider does not offer features which might influence their vendors’ pricing decisions based on aggregated data about similar products offered by other vendors on the platform. Similarly, when automating pricing decisions on behalf of a third-party vendor, the pricing should be personalised for the product itself based on factors such as customer feedback and metadata about the product.

Secondly, decision-making criteria should be transparently displayed at every opportunity. An emerging trend in machine learning is Explainable AI (see here: An Explanation of What, Why, and How of eXplainable AI (XAI) | by Bahador Khaleghi | Towards Data Science). Unlike most ML techniques, Explainable AI exposes which criteria led to a particular decision being reached. This is powerful since it enables significantly greater transparency of decision-making. Platform providers should consider favour more transparent algorithms over alternatives to provide consumers with clear information about which factors are used in making decisions that affect their choices.

Thirdly, design from the customers’ perspective. One of the major trends which has emerged from the digital revolution is that customers have much hire expectations with respect to their experience of a product. Designing your services to ensure they are as slick and customer-friendly as possible is just good business sense. However, a fortunate side-effect of this should be that by offering slick customer experiences, you will also be lowering the barrier to adopt your service and therefore avoid the risk that smaller customers will be less able to integrate with you.

Finally, decision-making algorithms should exclude criteria which would solely benefit the platform provider (for example, ranking products based on the profitability to the platform). Platform providers make a simple promise to customers: to provide them with the best available products or services based on price, quality, or a combination of both. As such, the criteria that platforms use when making decisions that impact the customers’ choices or wallet should be based exclusively on the interests of that customer.

The CMA is continuing to develop its guidance and thinking with respect to algorithms and competition law. These design principles will hopefully go a long way to mitigate the competition risks for digital businesses. We would, however, encourage business leaders to seek professional legal advice wherever they have concerns about competition risks.

Algorithms and competition law is one of our expert articles, created by our team who have more than 30 years experience in the digital environment, including heavily regulated industries and a wealth legal expertise. At Volanto we are on a mission to democratise digital. Contact Volanto here to see how we can unleash your potential.

Read more about AI: 

At Volanto we love seeing tech used to make lives better:

- Bowel Cancer blood test developed by UK firm 

- Multiple cancer blood test developed 

Read more about digital transformation and tech: 

- Women in Tech

- Creating company culture in the tech world

Sign up to our newsletter for updates on algorithms and competition law, as well as other legal and regulatory issues that often affect tech business. At Volanto we are digital growth and tech experts. Contact us now to see how we can unleash your potential. Algorithms and competition law are one area we can assist with. Algorithms and competition law are a complex area - make sure you take the time to understand the legal implications before utilising an algorithm. 

Add new comment

Restricted HTML

  • Allowed HTML tags: <a href hreflang> <em> <strong> <cite> <blockquote cite> <code> <ul type> <ol start type> <li> <dl> <dt> <dd> <h2 id> <h3 id> <h4 id> <h5 id> <h6 id>
  • Lines and paragraphs break automatically.
  • Web page addresses and email addresses turn into links automatically.