How is the advertising and tracking services ecosystem of smart TVs organized?

Add to my custom PDF

The TV is Smart and Full of Trackers: Towards Understanding the Smart TV Advertising and Tracking Ecosystem

Smart TV adoption has grown steadily over the last few years, with more than a third of US households owning at least one as of 2018. Most applications on smart TVs are supported by advertising. Despite the increasing popularity of smart TVs, their advertising and tracking services are not well understood by users, researchers or regulators.

Varmarken et al. conducted a large-scale study of the smart TV advertising and tracking ecosystem. They monitored the network traffic from 41 homes in a major US city, collecting the data generated by 57 smart TVs over 3 weeks. They then experimented with the ‘Roku’ and ‘Amazon Fire’ TV platforms to obtain more granular information as to which specific apps were generating the traffic. Using an automated testing system simulating user interaction, they collected the network traffic of approximately 1,000 of the most popular apps on each platform. The authors also evaluated the effectiveness of four popular systems that block advertising and tracking traffic from home devices, which are based on DNS blacklists.

Results indicate that smart TVs generate a substantial amount of advertising and tracking traffic. The Roku and Fire TV ecosystems differ substantially. For example, SpotX is a relatively large player on Roku, but is almost absent from Fire TV. In contrast, Facebook has almost zero presence on Roku, but has a reasonable foothold on Fire TV. The exception is Alphabet Inc., which has a strong presence on both platforms. Even apps present on both platforms show little overlap in where they are sending their data to, which further highlights the distinct nature of the advertising and tracking services between these platforms. All four tested blocking solutions were ineffective in filtering advertising and tracking traffic. The authors suggest a few ways in which they could be improved. For instance, they observed that the more apps contact a single destination, the more likely it is for that destination to be an advertising or tracking service. They also note that some obvious domains containing keywords such as ‚‘ads’ and ‘tracking’ were not blocked by any of the solutions.

The smart TV advertising and tracking services ecosystem appears to be fragmented. It is also different from the mobile ecosystem, which is better understood. More could be done to increase the effectiveness of blocking tools. The authors have made their datasets publicly available, and plan to also share their analysis tools.

The smart TV advertising and tracking services ecosystem appears to be fragmented. Blocking solutions are ineffective.