I'm not grokking your 'reporting site' reference.

There is reporting origin that must be identical across Source and Trigger registration responses, correct? IOW, https://ara.adtech.example or https://ara.advertiser.example on both sides.

Destination site, the location where Triggers are registered whose eTLD+1 must match prior source registrations by the same reporter.

And source site, the eTLD+1 where sources are registered.

If by {source site, reporting site} you meant {source site, destination site}, wouldn't that limit an advertiser to a single measurement partner (reporting origin)?

@dmdabbs if I understand correctly, this is about restricting the number of origins a given site can use as reporting endpoints, on one source site. That is, in the following sequence:

1. a.adtech.example registers a source on publisher.com
2. b.adtech.example registers a source on publisher.com

(2) would be disallowed by this new rate-limit, because the reporting endpoint's site (adtech.example) has already had one origin register on publisher.com.

@csharrison thank you for the clarification!
I certainly understand the need to head adtech privacy budget attacks off at the pass.

One use case for which this might produce friction is a multi-brand CPG or other entity using the same origin for various 'product lines' conversions. The advertiser or their partner will need to use separate reporting origins for each in order to create separate 'attribution lanes' for each line. Those lines would no longer have use of linea.advertiser.com, brandb.advertiser.com &c.

One use case for which this might produce friction is a multi-brand CPG or other entity using the same origin for various 'product lines' conversions. The advertiser or their partner will need to use separate reporting origins for each in order to create separate 'attribution lanes' for each line. Those lines would no longer have use of linea.advertiser.com, brandb.advertiser.com &c.

Thanks for the feedback! We hoped to solve this in some way with the attribution filtering mechanism, but this doesn't quite support multiple independent lanes like multiple reporting origins does. The filtering mechanism just lets you selectively ignore conversions that attribute to the wrong category/lane.

The problem with wholey supporting this use-case is that unfortunately it can lead to unintuitive privacy violations.