Get to know the new Topics API for Privacy Sandbox – Google
On January 25th, a year after the introduction of FLEDGE in place of TURTLEDOVE, Google introduced Topics API in place of FLoC. It is a proposal for interest-based advertising, which according to Google represents a significant privacy improvement over the previous iteration by incorporating industry feedback.
Google describes Topics API as more marketer-friendly, as it will provide a handful of browser-determined topics, such as “Fitness,” that represent user top interests for a given week. They will be selected entirely on the user device, and sites visited by the user will then receive only one topic per each of the last three weeks. Users will also be allowed to remove uninteresting topics that were assigned to them or opt out of this feature completely.
The company also states that due to its design and a curated taxonomy of topics, sensitive categories will be excluded. With the proposal, Google aims to provide websites with a powerful tool for ad personalization, which will not involve covert tracking techniques.
Together with this blog post, Google published an overview of Topics API on the Privacy Sandbox website and a technical explainer. The full, final specification will depend on the industry’s feedback. RTB House has already published its first impressions on Google’s news.
FLEDGE API developer guide – Google
Two days after publishing Topics API, Google also released long-awaited information about FLEDGE’s scope for the upcoming first Origin Trials. This proposal aims to satisfy remarketing and custom audience use cases.
Google starts with showing how and where FLEDGE API can be tested, pointing to the demo website and feature flags settings in Chrome. As the first Origin Trials will include a limited version of FLEDGE, it also covers what will be included at this stage:
Interest groups, stored by the browser
On-device bidding by buyers based on the interest groups
On-device ad selection by sellers
Ad rendering in a limited version of Fenced Frames
Later in the entry, Google explains the full FLEDGE process step-by-step, starting from the interest group definition, through bidding, auctions, ad rendering, and reporting. Google also covers how FLEDGE differs from the original iteration, TURTLEDOVE. One of the key differences mentioned is the addition of Product-level and Outcome-based TURTLEDOVE from RTB House, which improved the personalization capabilities and anonymity model.
However, the blog post does not mention that the Origin Trials specification won’t apply the k-anonymity model, making it possible to build interest groups at a user-level. It also does not inform the industry what will be the share of traffic with the FLEDGE API included. Lastly, the detailed timeline for the tests is also missing. RTB House has also published its first impressions on the FLEDGE specification.
In bad news for US cloud services, Austrian website’s use of Google Analytics found to breach GDPR – TechCrunch
Austria’s data protection authority complained that a website using Google Analytics was sending Europans’ IP addresses to Google located in the US. In this case, it is mostly about the improper use of the IP address anonymization function, but it touches on the general rule of sending IP addresses to the United States.
The DPA’s view is that sending IP addresses to the US is a violation of the GDPR, as it can serve as an element, which upon proper use, can lead to identifying an individual. If the data was protected properly, then the authority would not intervene. However, due to US surveillance law, it could be accessible by US intelligence agencies.
This decision falls under the umbrella of a larger EU-US data protection conflict, in which the EU requires the data of Europeans be protected abroad at least in the same way as in the EU, while the US requires access to data due to surveillance acts. Up until 2020, there was EU-US Privacy Shield in place, which regulated data transfers, but it was cancelled by the EU.
According to Max Schrems, key privacy advocate, this decision says that companies can’t use US cloud services in Europe anymore. He adds that in the long run, there can either be implemented new privacy protections in the US, or the products for the US and the EU will be fully separated.
If personalized advertising is banned, who bears the cost? – Mobile Dev Memo
Eric Seufert from Mobile Dev Memo starts his article with the summary of recent regulatory movements in the US, where the Bill to Ban Surveillance Advertising was proposed, and in Europe, where the Digital Services Act was voted in recently. Both of these propose restrictions on personalised advertising.
One of the major issues raised by the author is that the US bill does not take into account the user choice. In his view, it should be the user’s decision whether to pass on their data and benefit from the content and ad personalization.
The author also provides proof that banning targeted advertising would actually hurt the digital economy, unlike some of the politicians’ claims. As one example, he brings up Facebook analysis prior to ATT introduction on iOS (limitation on the access to IDFA), which claimed that ad personalization contributes to 50% of the CPMs. Another analysis was from the UK’s Competition and Markets Authority (CMA), which claimed that banning cookies would result in a 70% drop in publisher revenue.
Lastly, he emphasises that a group of publishers submitted a complaint to the EU to stop Google from switching off cookies, as it would harm their businesses. The author proposes two key areas to focus on in the further discussion on this topic:
To recognize the privacy/utility trade-off, which impacts how we look at the advantages of personalization for the overall user experience.
To recognize the economic impact of banning ads personalization, as not only large adtech players, but also advertisers and publishers will be affected by this change.
If you have any questions, comments or issues, or you’re interested in meeting with us, please get in touch.
