In its Google Search investigation, the CMA defined the relevant digital activity as comprising:

i. “a service that searches the world wide web, and can draw on other sources, to return information on any subject” – the general search function; and 

ii. “a service that enables advertising to users of general search” – the search advertising function. 

The CMA grouped these two digital activities into a single digital activity: “general search services.”  It considered the various services to constitute a “two-sided platform”, noting that they are “carried out in combination with each other to fulfil the specific purpose of providing a search engine.” 

Within these functions, the relevant products included, amongst others: (i) Google Search, and all information returned by it on the results page (including the new “AI Mode”, ‘News’ tab and ‘Top Stories’ carousel); (ii) AdSense (a tool used by publishers for displaying ads to website visitors) and (iii) Google Ads (a tool used by advertisers to purchase digital advertising²).

With generative AI chatbots increasingly being used by consumers when searching for information online, and general search engines increasingly incorporating AI features, the CMA examined the extent to which generative AI might fall within the scope of the relevant digital activities. This is a particularly pertinent question given the CMA’s strong interest in monitoring the market power of AI companies.

Google explained that it has used “machine learning and AI technology in Search for at least a decade.” The CMA therefore concluded, with Google’s agreement, that Google Search’s new AI Overview and AI Mode features are “deeply embedded within the Search infrastructure” and fall within the scope of Google’s general search services digital activity. 

However, that is not to say that all generative AI will be grouped into this designation in the future.  Notably, the CMA considered that Google’s Gemini AI assistant falls outside the scope of the relevant search activities, on the basis that Google Search and the results generated by its underlying infrastructure comprise only one of many inputs that go into Gemini’s response to a prompt. The CMA therefore considered that Gemini should instead be viewed as a “user” of Google’s general search. 

The CMA nevertheless recognised that, whilst Gemini does not “at this point in time” fall within the scope of the relevant digital activity, it is possible that if there are material changes to the way Gemini is offered and used, it may in future end up within the scope of Google’s search function for the purposes of the SMS designation.  

The above approach acknowledges the evolutionary nature of an SMS designation under the DMCCA.  It is both possible and probable that digital activities will evolve during a designation period, and SMS firms must self-assess whether new/amended products fit within an existing designation. The DMCCA guidance therefore notes that effective SMS monitoring will include an ongoing assessment of whether it is necessary to revoke a designation or replace it with a new designation applying to a similar or connected activity. This will become particularly relevant where we start to witness digital activities, such as Gemini, evolve from their position at the time of an original designation.  

As set out in the CMA’s decision notice, a “further SMS investigation” may take place at any point during a designation period to determine whether to further designate Google in relation to any additional activity or function which it considers similar or connected to the general search services. 

Comparing the Proposed Decision and Final Decision in the round, it is evident that nearly all of the CMA’s provisional views remained the same. However, there are a few instances where the Final Decision indicates a change of understanding of the digital activities from that set out in the Provisional Decision. For example, Google was successfully able to argue that its syndicated search services (allowing third party publishers to incorporate Google Search into their own websites) should only fall within scope of the designated activities where they are configured to search the whole web, and not just specific sources. This suggests that, where a potential SMS firm is faced with overwhelming evidence of SEMP, there is merit in focusing its efforts on shaping the perimeter of the anticipated SMS designation.

In its Final Decisions, the CMA concluded that both Apple and Google hold SMS in various mobile-related digital activities, and opted to combine them into a single digital activity – a “mobile platform” – on the basis that those “individual digital activities together form an integrated package of complementary services and content”. The CMA also noted that mobile platforms are two-sided: connecting mobile content providers (i.e. app developers and website publishers) with users. 

The CMA considered that each company’s mobile platform was comprised of substantially the same (but not identical) digital activities:

• Apple’s mobile platform was found to comprise its smartphone operating system (iOS); tablet operating system (iPadOS); native app distribution platform (App Store); and mobile browser and browser engine (Safari and WebKit) on mobile devices.
• Google’s mobile platform was found to comprise its mobile operating system (Android); native app distribution (Play Store); and mobile browser and browser engine (Chrome and Blink) on mobile devices.

Notably, the CMA’s “mobile platform” definition does not include either Apple’s or Google’s hardware, and this influenced the CMA’s terminology during the course of its investigations. In particular, the initial investigation launch referred to mobile “ecosystems”, but the CMA moved away from this descriptor on the basis that a mobile “ecosystem” also covers hardware, as well as the apps that run on the devices, and those were outside the scope of the designations.

SEMP is distinct from the traditional concept of market dominance under Chapter II of the Competition Act 1998, and bespoke to the DMCR. Although both “involve the assessment of market power and may often draw upon similar factors”, the CMA emphasised that SEMP need not be construed the same way as dominance, and the scope of the assessment is different. In particular, the concept of a “digital activity” in a SEMP assessment is intended to be a more flexible frame of reference than a product “market” in a dominance assessment. The CMA also rejected Google’s submission that SEMP sets a higher standard than dominance. This is despite the fact that, unlike for dominance, there is an express forward-looking element to the SEMP test: the CMA must conduct a forward-looking assessment, considering relevant likely developments over at least a five-year period. 

In line with its digital markets guidance, the CMA undertook case-specific SEMP assessments for Apple and Google respectively, citing the same key factors as being relevant in both cases:

• the level and stability of shares of supply;

• the number and strength of competitive constraints to the SMS firms;

• profitability levels; and 

• levels of customer switching. 

Ultimately, the CMA concluded that Apple and Google have SEMP in respect of their mobile platforms.  In reaching this conclusion, the CMA referred to both firms’ large and stable groups of end-users. The CMA stated that it does not expect Apple’s share of supply of the market to change significantly over the next five years, and referred to Apple’s restrictions on alternative app distribution and browser engines, and its consistently high profits. As for Google, the CMA examined the possibility of a new mobile platform for use in third-party mobile devices emerging, and concluded that it did not expect any such entry over the next five years given the presence of high barriers to entry/expansion.

In its forward-looking assessment, the CMA considered whether the growth of generative AI could impact Apple and Google’s market power. Whilst it noted that AI is likely to change the way users access and interact with content on their mobile devices, this is unlikely  to adversely impact Apple’s or Google’s mobile platforms. Indeed, the CMA found that both Apple and Google were likely to incorporate AI into their own offerings, meaning that barriers to competition in mobile platforms are likely to persist. 

Interestingly, the CMA found that Google’s mobile platform only imposes a limited competitive constraint on Apple’s (and vice versa), on the basis that end-users do not consider them to be close substitutes. Rather, end users think of them as distinct offerings and rarely consider switching, whether due to brand loyalty or the high barriers associated with switching.³ Competitors, including significant players such as Amazon, have been unable to overcome the high barriers to entry/expansion – particularly, network effects in relation to the prevalence of native apps – by which both firms’ mobile platforms are protected.

A side-by-side comparison of the Proposed and Final Mobile Platforms Decisions reveals significant changes between the two reports, with the CMA adding substantially more information and analysis to support its conclusions. As with the Google Search Final Decision, the vast majority of those conclusions remained the same, but both Apple and Google were successful in arguing for certain revisions to the scope of the products falling within the designated activities. These included, for example, clarification of which Android “middleware” products are captured, and similarly which of Apple’s interconnectivity APIs are included in the description of its iOS operating system.