You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository was archived by the owner on Oct 27, 2025. It is now read-only.
Repository navigation
This repository was archived by the owner on Oct 27, 2025. It is now read-only.
Protected Audience API and Dynamic Creative Optimization #201
For context, in the current programmatic retargeting using 3P cookies I am able to run Dynamic Creative Optimization (DCO) campaigns using a combination of Campaign Manager 360 (CM360) Floodlight variable and Google Studio (Product Catalog). The way this works, when a user browse a product on the advertiser's website the product ID is sent to CM360 via Floodlight variable using the user's 3P cookie. This means that I can create a single interest group say "product viewed" and the right product image is shown in the Display ad.
In PAAPI there is the adComponents attribute and the documentation says:
"Null or a list of interest group ad. Contains various ad components (or "products") that can be used to construct ads composed of multiple pieces — a top-level ad template "container" which includes some slots that can be filled in with specific "products"."
In this scenario:
When user A visits Advertiser website he is particularly interested in shoes. I create an interestGroup named shoes with the following ads and adComponents (please ignore the JSON formatting for now):
Thanks for the question @TheTamFamily. Could you please add your affiliation to your GitHub profile?
adComponents are meant for ads composed of multiple components, e.g. one display ad that advertises both running shoes and hiking boots. You can see the original proposal with an example ad here.
In the case you described where you want to display a running shoes ad to user A and a hiking boots ad to user B, adComponents probably isn't particularly useful. Instead you can just have one interest group with something like:
Thanks Paul. Your suggestion would suffice for a trial however, when this is in production i.e. when we run always on Display campaigns, this would not scale for Advertisers with a huge product catalog where there are literally thousands of product categories. This would mean hundreds of URLs pointing to each creative.
Ad composition and rendering currently is done by the Ad server and at render time, there is a single Ad creative that is instructed to show a particular image. If we have to hard code every Ad creative it means creating many line items.
I would like to propose the ability to pass from the key/value server the ad image (whether an jpg, png etc ...) to the fenced frame at render time. Is there any reason why this would break k-anon or any privacy rights of the individual.
Activity
JensenPaul commentedon Jan 9, 2024
Thanks for the question @TheTamFamily. Could you please add your affiliation to your GitHub profile?
adComponentsare meant for ads composed of multiple components, e.g. one display ad that advertises both running shoes and hiking boots. You can see the original proposal with an example ad here.In the case you described where you want to display a running shoes ad to user A and a hiking boots ad to user B,
adComponentsprobably isn't particularly useful. Instead you can just have one interest group with something like:ads: [ { renderUrl: 'https://www.examle.com/ig1/runningshoes.html', }, { renderUrl: 'https://www.examle.com/ig1/hikingboots.html', } ]or you could have separate interest groups for each user, that each contains only the specific ad you intend to show that particular user.
TheTamFamily commentedon Jan 12, 2024
Thanks Paul. Your suggestion would suffice for a trial however, when this is in production i.e. when we run always on Display campaigns, this would not scale for Advertisers with a huge product catalog where there are literally thousands of product categories. This would mean hundreds of URLs pointing to each creative.
Ad composition and rendering currently is done by the Ad server and at render time, there is a single Ad creative that is instructed to show a particular image. If we have to hard code every Ad creative it means creating many line items.
I would like to propose the ability to pass from the key/value server the ad image (whether an jpg, png etc ...) to the fenced frame at render time. Is there any reason why this would break k-anon or any privacy rights of the individual.
Thanks,
David