michaelkleber / privacy-model

A Potential Privacy Model for the Web: Sharding Web Identity
Other
189 stars 31 forks source link

Request for definition: "effective advertising" #6

Open ehsan opened 5 years ago

ehsan commented 5 years ago

What is "effective advertising" as is referred to in the text?

This document describes a way the web could potentially work that would not require cross-site tracking, but would still let publishers support themselves with effective advertising.

It seems that this text is presupposing that targeted advertising based on a user behavioral profile is effective/desirable and is trying to come up with a model to maintain that capability as an explicit goal, without really providing facts demonstrating that model of advertising is really core to the survival of the Web in the long term and cannot be done without.

As a strawman, is what's commonly known as "contextual advertising" (displaying advertisements based on the contents/URL of the page the user is visiting only) considered ineffective in this text?

michaelkleber commented 5 years ago

Excellent question! And I think one that needs to be answered as part of the discussion with the ads industry as well as the web development community. But this work was partly inspired by data behind the Google Ads blog post https://www.blog.google/products/ads/next-steps-transparency-choice-control/

Based on an analysis of a randomly selected fraction of traffic on each of the 500 largest Google Ad Manager publishers globally over the last three months, we evaluated how the presence of a cookie affected programmatic revenue. Traffic for which there was no cookie present yielded an average of 52 percent less revenue for the publisher than traffic for which there was a cookie present. Lower revenue for traffic without a cookie was consistent for publishers across verticals—and was especially notable for publishers in the news vertical. For the news publishers in the studied group, traffic for which there was no cookie present yielded an average of 62 percent less revenue than traffic for which there was a cookie present.

That experiment kept all "contextual advertising" in place. So whatever "effective advertising" is, I want it to offer the display-ads-supported web sites of the world a future that's better than "you lose half to two-thirds of your income".

ehsan commented 5 years ago

And I think one that needs to be answered as part of the discussion with the ads industry as well as the web development community.

Why are those two groups the right set of constituents to define what effective advertising means for the purposes of this proposal? For example, should the voice and interest of users be taken into account?

But this work was partly inspired by data behind the Google Ads blog post https://www.blog.google/products/ads/next-steps-transparency-choice-control/

I wish the actual data and the methodology for its collection was published as well, because then perhaps its relationship to the current proposal may have been more clear. It's possible that you've seen more information about this study that doesn't exist in the public blog post, but in its written form I'm afraid it doesn't support any of the conclusions you're drawing from it.

That the revenue of the "500 largest Google Ad Manager publishers" over three months for traffic without any cookie present was N percent less/more than traffic with cookies is an interesting anecdote, but it doesn't have any relationship to what this document is discussing, at least none that I can see.

Let's take a step back and think about what this document is discussing: "A Potential Privacy Model for the Web". It isn't titled "A Look at the Current Landscape of the Impact of Cookies on Advertising". It seems that overall the proposal is declaring that a global shared state that allows entities to access global shared IDs isn't OK, it promotes similar data/communications partitioning that is already implemented in WebKit and Gecko, and it calls for third-parties to have access to first-party data + a small surface of cross-site data for specific use cases.

If there is a desire to assess the impact of this proposal over the economics of publishers, it is prudent to actually do the work that is necessary to assess the economics of the model that is being proposed in this text to the status quo of the world. Comparing publisher revenue over "traffic with cookies" vs "traffic without cookies" is an apples vs. oranges comparison, and to be honest it really feels like an attempt at grabbing tech news headlines (given the promotion it's receiving on the official company blogs).

Apologies if I'm appearing a bit harsh, but talking about revenue numbers from the lack of cookies in today's world very commonly comes up in these discussions when strong privacy proposals are being discussed, so I wanted to address this issue head on without dancing around it.

That experiment kept all "contextual advertising" in place. So whatever "effective advertising" is, I want it to offer the display-ads-supported web sites of the world a future that's better than "you lose half to two-thirds of your income".

To address the last paragraph of your post specifically, it supports no such conclusions. In the current world there is a market for the user's data that has been extracted from them through covert surveillance without their consent (this is commonly referred to as "targeted advertisement"). So when you're saying "you lose half to two-thirds of your income" you're actually saying "in today's world there are advertisers who would pay a premium of up to 200% for access to the user data extracted through covert surveillance without consent". That is all.

This isn't surprising, nor news. And it isn't relevant to the question of how the market would shift when it won't be possible to bid on the user's data that has been extracted through covert surveillance without consent.