On 9 June, the European Commission released the preliminary report from its market study into Consumer IoT. The purpose of the study is “to gain a better understanding of the consumer IoT environment, the competitive landscape and developing trends in this nascent sector”, but the report quickly unwinds into a statement of concern that Amazon, Google and Apple will be the ‘gatekeepers’ of consumer IoT through their voice assistant services.
The Chatty world of IoT
The report’s starting point is that ‘interoperability’ is fundamental to the success of any consumer IoT device. Some consumer IoT devices can only fully work when paired with another device: e.g. while ‘wearables’, such as fitness trackers, can display some basic information, their limited connectivity and reduced processing and storage capability means that they need to be used with a smartphone.
Other consumer IoTs may have their own ‘native’ user interfaces, but as they are directed at a specific task, a user having to control each device individually would be a frustratingly fragmented experience.
Individual consumer IoT devices supplied by different manufacturers often need to ‘talk’ to each other: for example, smart speakers depend on interaction with third party music streaming services; smart shutters on integration with a weather app; and smart thermostats on connection with heating systems.
While the report notes that currently the smartphone is the main means by which users access consumer IoT devices, “voice assistants represent the fastest developing interface for users to access the web.” The report considers that, with the investment and time required to build a high functioning general voice assistant service, the Amazon, Google and Apple services are likely to continue to be the primary platforms globally: while Google Assistant and Siri function in 20 to 30 languages, Samsung’s Bixby lags far behind in 8 languages.
There are two ways in which a third party consumer IoT can interconnect with a general voice assistant service. First, there is the ‘works with’ approach which allow manufacturers to make their smart home devices controllable through a voice assistant embodied on another device (i.e. smart speaker) or app support (i.e. smartphone application). The smart home device manufacturer uses the voice assistant provider’s documentation on API functionality to develop a voice application. The voice assistant provider then tests the developed integration to ensure a good user experience. If certification is granted, manufacturers can use “works with” logos or badges in their packaging or for online marketing.
Second, there is the ‘built in certification’ process which allows third party smart device manufacturers to support Amazon or Google’s cloud-based voice assistants on their devices and gain the “Alexa built-in” or “Google Assistant built-in” badges. Apple, on the contrary, does not offer built-in certifications, as Siri can only be built-in on Apple’s own devices.
While ‘works with’ allows the user to control the consumer IoT from the voice assistant ‘box’, the ‘built-in certification’ allows the user not only to control the IoT device, but to access the other functions of the voice assistant through the consumer IoT device: e.g. asking you smart fridge to search the internet for a recipe.
Currently, the report estimates that about two thirds of consumer IoT devices use the ‘works with’ approach and a third have ‘built-in’ certification.
Competition concerns arising from interoperability
The report says that the ‘gatekeeper’ role of the three big voice assistant platforms raises several competition concerns.
First, there are concerns about anti-competitive conduct through control of standards:
“Interoperability requires technical and business engagement among companies active in the consumer IoT sector in order to achieve meaningful integration between the different elements of an IoT ecosystem. In practice, however, such integration processes are largely determined by the presence of a few providers of leading proprietary voice assistants and operating systems relevant for the consumer IoT sector. These companies are able to determine independently the requirements to achieve interoperability with their proprietary technology through unilaterally governed terms and conditions, technical requirements and certification processes. By unilaterally governing the interoperability and integration processes, some respondents to the Sector Inquiry indicate that the leading voice assistant providers may also be able to limit the functionalities of third-party smart devices and consumer IoT services, compared to their own, by imposing technical constraints, such as limited APIs.”
Second, there is a risk of ‘disintermediation’ of the consumer IoT manufacturers by the general voice assistant service providers. Once integration with a third party user interface has been achieved, the brand’s own user interface usually becomes redundant and the consumer IoT manufacturer loses the direct connection to the customer. Further, the report says that as a user will often need to go through the general voice assistant’s onboarding process to register a new consumer IoT device, the manufacturer can lose its ability to control the user experience of its device.
Third, in relation to data, because voice assistants could become a central ‘node’ of the consumer IoT, “voice assistant providers accumulate large amounts of data and to not only control data flows and user relationships, but also to leverage these advantages into adjacent markets, i.e. the provision of other consumer IoT products and services.” The report says that not having access to data can raise barriers to new entrants on the voice assistant market and hinder the development of smaller competitors on that market.
The report canvasses whether common industry standards or mandating use of non-proprietary technology would help address these competitive risks. The report notes the industry itself, including with the active participation of Big Tech, are engaged in standardisation efforts. In December 2019, Amazon, Apple, Google and the Zigbee Alliance joined together to promote a new working group, the Connected Home over IP (internet protocol) or CHIP Project, to develop a new, royalty-free connectivity standard to increase compatibility among smart home products. Today, the CHIP Project has over 145 active member companies.
However, while the report says these efforts are making a difference at the basic connectivity level, “formal standards are currently not in a position to effectively compete with proprietary technologies of the leading providers of operating systems and voice assistants for other types of technologies such as device definitions, application layers and user interfaces.”
But the report also acknowledges that consumer IoT is such dynamic and heterogeneous technology environment that there would be a challenge in standard setting. As many consumer IoT manufacturers are SMEs, requiring adherence to a welter of new or rationalised standards might only add to the costs and challenges faced by them.
The report also noted that many consumer IoT manufacturers say they can work with proprietary standards provided there was enough transparency and an assurance of non-discriminatory application. The report notes that Amazon’s Voice Interoperability Initiative “is also seen as having the potential to bring together a broad user-base and grow into a leading proprietary technology (de facto standard), as a result of the interest of many industry players and Consumers to allow for a swift interaction with more than one voice assistant on the same device.”
Privacy concerns from interoperability
The report notes that interoperability means that voice assistants, consumer IoT services and smart devices collect and exchange significant amounts of data: millions of interaction data points and billions of logs monthly. For example, a smart washing machine may collect data on the washing cycle chosen and how often the user chooses this setting. A smart coffee maker may collect data on the settings selected when brewing coffee, collecting data on the preferred beverages and times for making coffee. An electric smart toothbrush can collect data on the pressure applied when brushing and the time taken.
The report notes that some consumer IoT ‘self-limit’, as it were, for privacy because the data processing is mainly done in the device itself (a form of edge computing): they do this because it allows quicker response times, avoids the problems of patchy network connectivity, and saves battery power because transmitting and receiving data to the cloud requires more energy than onboard processing. However, even so, the report says that most respondents said that they usually collected and stored consumer data in large data lakes.
The report noted that most respondents said that data was used for functional purposes between third party devices and internally within their own businesses (e.g. generating suggested content for a viewer. However, the report says that the consumer IoT sector has “unlocked new advertising space possibilities”. In particular, some types of smart home devices can show advertisements through displays (e.g. a smart fridge incorporating a display that might advertise food products). Likewise, voice assistants can serve digital audio advertisements.
The reports says that smaller consumer IoT manufacturers are concerned that, with disintermediation, it is the voice assistant providers not them who will monetise the data which is collected through their devices.
The report notes that these monetisation opportunities are not yet well developed, and will have to comply with the EU regulatory limitations on data use.
Where to from here?
The EC is currently consulting on the report. The report can be seen as another shot in the gathering global battle between Big Tech and regulators and policymakers:
Whether or not you agree with the report’s underlying thesis that ‘Big Tech must be reined in’ angle, it is a great backgrounder of the emerging world of consumer IoT.