User needs research

[edent]

There doesn't seem to be any evidence of user research. For example:

The type of ad targeting we propose supporting can be of great value to people browsing the web, who often prefer ads for things they are interested in

Which users have you spoken to, or what research have you done, to support this?

People who like ads that remind them of sites they're interested in can keep seeing those sorts of ads.

What evidence is there that such users exist?

There's a lot of great technical content here. But before designing an API, it would be helpful to do user research to see what it is that users want. Is there a desire for individuals to see why they've been targeted? If blocking specific advertisers is a user need, why don't existing tools meet this demand?

[michaelkleber]

For discussion about "people browsing the web, who often prefer ads for things they are interested in", check out the Google A/B experiment from last year that observed the impact of dropping 3p cookies. Most of it is about money, and indeed that's the context in which I linked to the paper in the TURTLEDOVE Introduction section. But their "Additional Reflections" section includes this bit:

Another observation from the randomized experiment: users expressed greater dissatisfaction with non-personalized ads because they were not interested in what the ads were showing them. Users can choose to stop seeing an ad by clicking on an “X” that appears on a display ad to close the ad. We saw a 21% increase in user clicks to close an ad by the treatment group (who encountered non-personalized ads). When prompted with a list of reasons why they wanted to stop seeing an ad, there was a 21% increase in user clicks on the reason “Not interested in this ad” and a 29% increase in user clicks on the reason “Seen this ad multiple times”.

Of course that's behavioral data, not UX research. A Google search for Do users prefer personalized ads shows links to lots of people claiming data on that question, and of course they have a wide range of numbers and sometimes contradictory conclusions.

So if the browser takes on the role described in TURTLEDOVE, we will absolutely need some UX research on how to present it to our users. The User Interface Controls section of the Explainer has a few preliminary thoughts about the browser could offer, once it's directly involved in how interest groups are stored and used.

Note, of course, that the UX choices are the domain of each individual browser, not part of any standards process.

[edent]

I think you're putting the cart before the horse.

For any new standard like this, we need to see evidence that there is a user need - or user desire - for it.

This proposal jumps from "I've had a great idea" straight to "and here's how to implement it" without asking "do the users want this?"

I suggest halting further development until you can demonstrate that this idea would be helpful to real users.

[michaelkleber]

Ah sorry, I misunderstood. From that point of view, the user desire that we're serving here is privacy. The whole Privacy Sandbox is a suite of efforts to make the web dramatically more private than today, without breaking lots of things.

[ascentitall]

I agree with @edent in that users are confused, scared, and unaware of what tracking is and how it works. Instead of focusing on transparency as a first phase, we have jumped to the conclusion that all tracking is bad and tracking must be eliminated. When we know that tracking injects significant value into the ecosystem, between marketers and creators of free content. We have repeatedly seen this in decreased revenue to publishers when cookies are absent. What we do know is that the majority of users prefer free content over paying for content: https://www.networkadvertising.org/blog-entry/nai-consumer-survey-privacy-and-digital-advertising/

[michaelkleber]

@ascentitall This is a proposal about how to continue to inject significant value into the ecosystem, without the need for the widespread cross-site tracking that it's built on today. So I don't think what you're saying is a relevant complaint about TURTLEDOVE.

I quite agree that transparency is a valuable goal. Check out the User Interface Controls part of this Explainer for the kind of transparency that browsers could be in a position to offer here.

[edent]

I'd like to keep this discussion open. Have you spoken to any end-users about this proposal?

[TheMaskMaker]

Just following up on this thread; I agree an unbias user study would be helpful here. The difficulty would be explaining to the users what the change entails. The proposal does not prevent tracking; it changes how tracking works, and which powers have control of what and how it is limited.

An early user study determining the primary change that is definitively part of fledge, i.e. cohort tracking and the differences to the current system, would be useful. This is a drastic change the the web technology stack. A later user study explaining the roles and powers of different players and where trust lies in how the data is transmitted could also be useful once Fledge is fully fleshed out, though it may be difficult to explain the innerworkings of adtech.

As part of a bias test different groups could construct different prompts to see if there is a large difference in user delight or rejection. We'd also have to account for sample bias (who would be willing to take the survey, and are they less indifferent) etc.