- From: Bosko Milekic via GitHub <sysbot+gh@w3.org>
- Date: Sat, 12 Feb 2022 04:17:22 +0000
- To: public-patcg@w3.org
Would the [Advertising Use Cases document](https://github.com/w3c/web-advertising/blob/main/support_for_advertising_use_cases.md) be a good starting point to source use cases to evaluate considered proposals (or eventual extensions of considered proposals) against? It already attempts to exhaustively map out use cases to Chrome and Safari (and some of the older community) proposals. We could probably focus on the attribution related use cases, e.g.,: - Conversion Lift Measurement - Brand Lift Measurement - Click-through attribution - View-through attribution - Multi-touch attribution - Cross Browser / Cross Device Measurement - Multi-channel attribution - We may want to separately consider ML modelling specific use-cases, in particular _Conversion Rate (CVR) Model: P(conversion|click)_, _Post-view attribution: P(conversion|impression)_, and _ROAS optimization_. We can of course add additional considerations to the list. For example (borrowing from @bmayd and @ssanjay-saran): - Report latency anticipated constraints - Report frequency anticipated constraints - Anticipated energy consumption for user agents (e.g., mobile battery drain) - Ability to debug and/or tune by operators (probably applies more to LARk and ML use cases) - Fairness: for conversion measurement, are both the advertiser and publisher (or their representatives/vendors) able to independently verify/calculate reported results? Finally, I propose that we consider privacy and security separately to utility. I suspect that some proposals will fair better than others from a privacy and security standpoint, while on the utility front a different picture may emerge. -- GitHub Notification of comment by bmilekic Please view or discuss this issue at https://github.com/patcg/meetings/issues/25#issuecomment-1036984614 using your GitHub account -- Sent via github-notify-ml as configured in https://github.com/w3c/github-notify-ml-config
Received on Saturday, 12 February 2022 04:17:24 UTC