How much more functionality should the browser perform?

[ydennisy]

Hi!

I would like to begin by saying that I have long been a believer in the concept which you have termed "Person-driven advertising" in your explainer.

However, post the start of a heavy move to make advertising more private I am seeing lots of proposals for shifting more and more logic into the browser, which mean centralising more of that functionality since there are really only 3-4 browser providers and less engines.

So my question is broad and maybe not directly related to this specific proposal but I still feel is relevant, and that is how much more of this logic can we shift to the browser and away from "open web" apps and services which anyone is free to build and innovate on in a much quicker fashion than standards and committees?

To me it feels these are no longer standards and protocols, they are full fledged applications. Which draws the question why should applications be baked into the browser specs?

Look forward to your feedback :)

[dmarti]

In general it is important to balance the speed of innovation that can be done on individual web sites with the ability of in-browser code to be more closely aligned with the interests of the user. (The browser is the agent of the user, so browser code is incentivized to deliver advertising that is a net win for the user, ad impressions where value of incremental ad-supported content+economic signal > computing resources+annoyance+risk)

In the case of this specific proposal, some users have propensities to act on an ad that are not captured by conventional third-party tracking. For example, a user might choose to "follow" another user's ad topics. For example, a user who wants to begin developing new nutrition and exercise habits might choose to follow their coach or instructor's ad preferences.

From the web publisher point of view, having a browser mechanism to rate ad topics is also a helpful signal for optimizing a site and blocking ad topics if necessary. Today, without a browser signal it is hard for an independent site to see if a user is bouncing from a site because they didn't like the content or because they don't like ointment ads and they happened to get an ointment ad.

[ydennisy]

Thanks @dmarti for your reply.

I think that maybe my original point was not clear. I am all for the concept, and have been noodling this exact idea for a few years.

Where my question or concern comes from is what the browsers role here should be. In the past the browser has given building blocks for developers to use to build applications on top off. What is being suggested here and in many of the privacy sandbox style proposal is not really a building block in my eyes but crossing over into application space.

I see many problems with this approach, primarily the speed to innovation, divergence between browsers and centralisation of what is built. The key problem is that there are many interesting ideas outside of the W3C.

So, I think it would make sense to at least consider what the browser should provide and where to draw the line. Even for this proposal:

• creating a standard way for users to block or request ads of specific categories makes sense as a standard
• building functionality to follow other users, or specific embedding models etc does not seem like a standard/protocol

The browser is the agent of the user, so browser code is incentivized to deliver advertising that is a net win for the user

I do not agree with this statement, whilst this is the ideal scenario, we have seen with recent FLOC problems that this is not always the case in the eyes of some people. At the end of the day the user agent is owned by a company who has other incentives.