Preferential treatment in ranking: what is it?
There has been debate around the terms by which a large digital platform should treat its own offer relative to others’ offers, and to what extent preferential treatment for its own offers constitutes a concern. In 2017, the EC found such conduct to be abusive in the context of Google Shopping where it found that ‘Google positions and displays more favourably, in its general search results pages, its own comparison shopping service compared to competing comparison shopping services’.
A similar concept has found its way into the EC’s Digital Markets Act (‘DMA’) in the form of a prohibition of engaging in ‘more favourable treatment in ranking services’. The DMA builds upon the definitions of ‘ranking’ and ‘differentiated treatment’ in the Platform to Business (‘P2B’) regulation, by prohibiting such behaviour if engaged by gatekeeper platforms (while P2B only contained provisions on information sharing). In practice, this implies that there is a two-layer regulation. All platforms need to be transparent about their conditions of service, including terms applied in ranking (as per the P2B regulation), but only (very) large intermediaries are asked not to rank their own services or products more favourably than those offered by third parties.
The DMA refers to preferential treatment in the context of ranking, in Article 6d, by stating that a gatekeeper shall ‘refrain from treating more favourably in ranking services and products offered by the gatekeeper itself or by any third party belonging to the same undertaking compared to similar services or products of third party and apply fair and non-discriminatory conditions to such ranking’.
By the EC’s own ‘disclaimer’, obligations listed under Article 6 in the DMA are ‘susceptible of being further specified’. The above definition raises a number of questions which discussions around further specification would need to consider.
What is ‘preferential ranking’?
The DMA is not clear about how ‘treating more favourably in ranking’ is defined. The aim seems to be to set rules to be neutral vis-a-vis any third party and any gatekeeper-owned competing service. However, even a genuinely neutral algorithm may not always ensure a level playing field in the outcomes. For example, if a gatekeeper search engine owns a video-sharing platform and a user searches for a video-related query, the search engine could display results for the gatekeeper-owned video platform first, because that was more relevant to the query based on its algorithm. Would this conduct be prohibited under the DMA? If so, in an extreme case, should the product or service of the gatekeeper never be entitled to perform at the top of the ranking? Even if the conduct is not prohibited, would it lead to complaints from rivals, and if so, how would the EC respond?
In fact, many outcomes that the EC might consider to be ‘preferential treatment’ could be the genuine results of an objective algorithm. Therefore, there is an unresolved tension between a gatekeeper’s product or service performing better than that of a rival (according to some objective and transparent criteria) and creating a level playing field in ranking outcomes.
There is also a question of whether a truly objective algorithm can ever be developed. Ranking algorithms typically weight different alternatives across various criteria, according to some underlying principle such as relevance to the user. But identifying the criteria and the weights on those criteria will likely involve some level of choice that is outside the algorithm itself. How will the EC determine whether certain criteria should have been included or omitted? And how will be EC determine what the appropriate weights should be?
What is a ‘similar’ product or service?
The EC refers qualitatively to the concept of ‘similar’ products and services. The interpretation and application of this part of the definition, lacking further guidance, may be arbitrary. A ‘similar’ product or service could be interpreted as a substitute product or service. But how close a substitute can a product or service be before it is ‘similar’? One could potentially carry out a market definition exercise to identify products that are sufficiently similar. But it is well known that when products are differentiated, even products in the same market might be quite different from each other. At the other end of the spectrum, competition authorities often define markets very narrowly (using a largely functional characteristics-driven approach), and so there may be quite similar products outside the market.
These problems arise using market definition as a basis for determining ‘sufficient similarity’. Greater confusion would arise if one were to move away from the market definition standard.
The EC’s description of preferential ranking in the draft DMA legislation leaves significant questions unanswered. In its current form, this requirement on gatekeepers gives rise to material uncertainty and challenges for compliance. As such, additional clarity and specificity from the EC on the definition of this practice, and the application of the relevant tests, would be welcome.