Although there were some basic metrics published about the FLoC OT (e.g. number of cohorts, percent of cohorts deemed sensitive and dropped), there was little else published about whether the FLoC OT showed whether FLoC was useful for interest-based advertising or preserved privacy.
Such an analysis would be useful to help evaluate future proposals, such as Topics.
A colleague and I did our own post-mortem analysis using our implementation of FLoC and a proprietary dataset of user browsing histories:
https://arxiv.org/pdf/2201.13402.pdf
It would be useful for the research and web community for the Google Chrome team to publish such an analysis from the true OT data and FLoC implementation.
Although there were some basic metrics published about the FLoC OT (e.g. number of cohorts, percent of cohorts deemed sensitive and dropped), there was little else published about whether the FLoC OT showed whether FLoC was useful for interest-based advertising or preserved privacy.
Such an analysis would be useful to help evaluate future proposals, such as Topics.
A colleague and I did our own post-mortem analysis using our implementation of FLoC and a proprietary dataset of user browsing histories: https://arxiv.org/pdf/2201.13402.pdf
It would be useful for the research and web community for the Google Chrome team to publish such an analysis from the true OT data and FLoC implementation.