Programmatic Advertising: What Data Signals Could Google Use to Build Its Audience Segments? 🕵️

A while ago, we explored the mystery surrounding Google's audience segments in Display & Video 360 (DV360) and Google Ads:
https://greyfox-ads.com/articles/programmatique-le-mystere-des-segments-de-data-google/

One question remained unanswered: what kinds of user data might Google collect, and which signals could potentially contribute to its audience modelling?

We found an interesting clue. Not proof of Google's audience-segmentation recipe, but a fascinating glimpse into the data that tracking technologies may try to access.

What Is DuckDuckGo's App Tracking Protection?

You probably know DuckDuckGo. 🦆 It started as a privacy-focused search engine and now also offers a browser that competes with Chrome.

Its Android app includes a particularly interesting feature: App Tracking Protection.

The feature identifies and attempts to block certain third-party tracking requests from other Android apps. It also provides a dashboard showing:

• Which companies are associated with detected tracking attempts
• Which apps generate those attempts
• Which categories of personal or device information trackers may seek
• How many tracking attempts were blocked

Naturally, we had to try it. 🍿

What Did Our Seven-Day Android Test Reveal?

Over seven days, we used 19 apps on our test device.

According to DuckDuckGo's App Tracking Protection dashboard, it blocked an astonishing 27,579 tracking attempts. 🤯

Google-associated trackers appeared in 18 of the 19 apps in our sample. Netflix was the exception in this particular test.

Important context: this is a single-device observation, not a representative study of all Android apps. A blocked attempt is not necessarily a successful transfer of personal data, and repeated attempts do not necessarily represent different pieces of information being collected.

Which Data Signals Did DuckDuckGo Identify?

The dashboard associated Google-related tracking attempts with a broad range of potential data categories, including:

Identity and identifiers

• First name
• Last name
• Gender
• Email address
• Unique identifier
• Android advertising ID
• Cookies

Location and network

• Local IP address
• GPS coordinates
• City
• Postal code
• Country
• State or region
• Time zone
• Network connection type
• Mobile network operator

Device and technical information

• Device name
• Phone manufacturer
• Phone model
• System volume
• Screen dimensions
• CPU information
• Device orientation
• Available internal storage
• Screen resolution
• Battery level
• Charging status
• Headphone connection status
• Device uptime
• Total device memory
• Operating system version
• Operating system build or version number
• Language

App information

• App name
• App installation date
• App version

That's quite an ingredient list. 👀

Does This Mean Google Uses All These Signals for Advertising?

No — and this distinction is essential.

DuckDuckGo's dashboard can provide visibility into tracking activity and potential data categories. It does not establish that every listed field was actually transmitted, retained by Google, linked to a person, or used to build an audience segment in DV360 or Google Ads.

Some requests may relate to analytics, app functionality, security, measurement or other purposes rather than advertising. And the exact signals and modelling methods behind Google's audience products are not fully disclosed.

So, have we cracked Google's secret audience-segmentation formula? Not quite. 🔍

But we have a useful reminder of just how many signals can exist across the mobile app ecosystem — and why transparency about data collection and audience modelling matters.

What Should Programmatic Media Buyers Take Away?

When evaluating an audience segment, its estimated reach is only one part of the story. Media traders should also ask where the underlying signals come from, what consent and privacy controls apply, how the segment is modelled, and how its incremental value can be measured.

At GreyFox Ads, we're a French programmatic advertising company focused on building a higher-quality Open Web for advertisers.

Our Master Curated Deal brings together:

📡 Supply Path Optimization (SPO) with direct publisher connections
🚫 Exclusion of made-for-advertising (MFA) sites and subdomains
🛡️ Brand safety filters
📐 Optimization for meaningful on-screen ad size (screen share)
🎯 Audience activation across cookie, cookieless and consent-dependent environments, subject to applicable privacy requirements
👌 Simple activation for programmatic media traders
🪟 Transparent costs

Looking for a quality-focused programmatic partner?

📧 Contact GreyFox Ads to discuss your display and video advertising challenges.

Original LinkedIn post: https://lnkd.in/edtujbQy