An early stage idea to share: Private Lift Measurement

Hello all,

I just added a few documents to the github repo that I'd love for you all to look over. I'm happy to answer any questions about them in our meeting tomorrow.

In broad strokes, we are just trying to share some early ideas and start a conversation. These are not finalized proposals, and we aren't endorsing any kind of specific implementation at this point. We just want to try to participate in the broader brainstorming about how we can work together as an industry to design privacy-preserving ways of supporting legitimate advertising use-cases.

As an engineer, I tend to favor incremental approaches to big projects. Just take one step at a time. Migrating web advertising away from technologies like 3rd party cookies is going to be a big project, and it'll take many small steps to get there.

I think a good first step, would be for us to try to find a privacy-preserving approach to Lift measurement. As I say in the document, it's a good candidate because it is by nature an aggregate measurement. This should reduce the risks related to personally identifiable data.

Nothing would make me happier than for us to find a good first step we can all take here, browsers and the ads industry, to introduce an API that both delivers value for advertisers, and that has really good privacy characteristics.

If we can take this first step, build trust, and demonstrate an ability to move forward together, I think the second step will be easier.

Here are the links to the documents for your consideration:

  *   Conceptual Overview: https://github.com/w3c/web-advertising/blob/master/private-lift-measurement-conceptual-overview.md
  *   Proposal for lift measurement when ads only run on a single first party website: https://github.com/w3c/web-advertising/blob/master/private-lift-measurement-first-party.md
  *   Proposal for lift measurement when ads run across many websites: https://github.com/w3c/web-advertising/blob/master/private-lift-measurement-third-party.md


-- Ben

Received on Wednesday, 9 October 2019 22:31:23 UTC