/
Navigation
Chronicles
Browse all articles
Explore
Semantic exploration
Research
Entity momentum
Nexus
Correlations & relationships
Story Arc
Topic evolution
Drift Map
Semantic trajectory animation
Posts
Analysis & commentary
Pulse API
Tech news intelligence API
Browse
Entities
Companies, people, products, technologies
Domains
Browse by publication source
Handles
Browse by social media handle
Detection
Concept Search
Semantic similarity search
High Impact Stories
Top coverage by position
Sentiment Analysis
Positive/negative coverage
Anomaly Detection
Unusual coverage patterns
Analysis
Rivalry Report
Compare two entities head-to-head
Semantic Pivots
Narrative discontinuities
Crisis Response
Event recovery patterns
Connected
Search: /
Command: ⌘K
Embeddings: large
TEXXR

Chronicles

The story behind the story

← → days · ↑ ↓ browse · Enter similar · o open

A database on Microsoft's Xandr containing 650K+ “audience segments” shows how data brokers classify consumers by surveilling their online and offline activity

A spreadsheet on ad platform Xandr's website revealed a massive collection of “audience segments” …

The Markup

Context & Ripple Effects

The segment list published on Xandr's ad platform extends a reporting arc The Markup has been building since late 2022, when it documented data brokers selling comprehensive voter profiles to campaigns. What was previously inferred from broker marketing materials is now visible as an actual product catalog: 650K+ named segments sitting in the open on a major platform's own site.

It also lands on a familiar failure mode for Microsoft specifically — researchers previously found 80M US household records exposed on a Microsoft cloud service, and the same pattern repeated with an ad agency's database of personal records. The company's ad-data infrastructure keeps turning out to be both vast and loosely guarded.

First-order effects

  • Advertisers and agencies buying through Xandr now have a readable map of exactly what consumer classifications the platform sells — including segments built from offline activity, not just browsing behavior.

Second-order effects

  • Rival platforms get pulled into comparison: The Markup's earlier investigation found Google's Display & Video 360 allowing sensitive-data targeting despite its own rules, so Microsoft now faces the same scrutiny Google did over whether its stated policies match its segment inventory.
  • Data brokers supplying these segments face reputational and compliance exposure — their taxonomies are no longer private sales collateral but public evidence of how consumers are classified.

Third-order effects

  • If platform segment catalogs keep leaking into public view, they become raw material for regulators assessing data-broker practices — turning ad-tech targeting from an opaque market into something auditable against privacy rules.
  • The recurring pattern of sensitive ad-targeting surfaces at Microsoft and Google suggests industry self-policing is structurally unreliable, strengthening the case for external enforcement rather than platform policy documents.

The trend: Ad-platform targeting inventories are being dragged from opaque sales catalogs into public and regulatory view, one exposed segment list at a time.

Discussion

  • @jonkeegan@mastodon.social Jon Keegan on mastodon
    My latest for @themarkup:  —  We reveal a spreadsheet of 650,000 advertising audience segments found on Xandr's (Microsoft) website full of thousands of sensitive segments relating to medical conditions, mental + reproductive health, sensitive locations and more. …
  • @wolfiechristl Wolfie Christl on x
    In this thread, I want to share some additional details about what the file from Xandr/Microsoft, which was reported yesterday (https://t.co/...), reveals about how hundreds of consumer data brokers trade personal information on billions of people at a global level.⬇️
  • @wolfiechristl Wolfie Christl on x
    Several data brokers including Epsilon (owned by French Publicis), Nielsen (UK), Neustar (TransUnion), Eyeota (Dun & Bradstreet), Oracle, Experian, LiveRamp sold lists of ‘union members’ via the Xandr data marketplace. ...which is highly problematic, not just under the GDPR. [ima…
  • @alexbeaurepaire @alexbeaurepaire on x
    Interesting piece from @themarkup about @xandr targeting techniques. It seems most people outside of the #AdTech world discover fire when they are shown targeting practices that are commonplace. The industry itself must be more educative with the public. https://themarkup.org/...
  • @chesterj1 Jeffrey Chester on x
    A key reason we opposed #Microsoft #Activision merger was #xandr—part of @Microsoft expanding commercial #surveillance for marketing. #Gaming & #Gamers will be a key place to collect & target via this data. #privacy #competition #consumers #FTC #health https://themarkup.org/...
  • @wolfiechristl Wolfie Christl on x
    Spent one more hour to extract a number of segments that are related to casino+online gambling. Several data brokers sold lists of gamblers in many different countries via Xandr, e.g. US/EU/UK, Brazil, Columbia, Chile, Japan. One company explicitly sold data on gambling addicts. …
  • @wolfiechristl Wolfie Christl on x
    Some more examples from my own analysis of the file. Xandr (re)sold lists of ‘disabled’ people and of the ‘poorest 10% of the nation’, payday loan prospects, people with ‘poor’ credit level, ‘poor’ income, ‘very poor’ financial stability, ‘chronically poor health caregivers’. [im…
  • @wolfiechristl Wolfie Christl on x
    The file is dated May 2021, but was available on the Xandr website until last week, when Microsoft who didn't respond to any media inquiry, took it offline. We don't know which segments they sell today. Microsoft still lists 63 data providers on the site. https://docs.xandr.com/.…
  • @eff @eff on x
    The effort by the online ad industry to closely target people, as in the segments file reviewed by @themarkup, constitutes “one of the greatest threats to data privacy” today, EFF's @Adam_D_Schwartz said. https://themarkup.org/...
  • @lawyerliz Elizabeth Wharton on x
    The labels given & tracked: violating #privacy & inaccurate .... From “Heavy Purchasers” of Pregnancy Tests to the Depression-Prone: We Found 650,000 Ways Advertisers Label You - The Markup https://themarkup.org/...
  • @johnnyryan Johnny Ryan on x
    Bravo @WolfieChristl. See coverage by The Markup's @jonkeegan here https://themarkup.org/... See 24mb excel sheet of 651,463 “audience segments”, including people's health conditions, religions, etc. https://web.archive.org/...
  • @tonyajoriley Tonya Riley on x
    Advertisers were offering segments that specifically targeted visitors of Congressional office buildings 👀https://themarkup.org/... [image]
  • @alantoner4 Alan Toner on x
    The ad industry surveils users behaviour on/offline to classify them in ‘segments’ which can be sold to advertisers. This process includes observation and inference on sensitive health, political and financial data. @WolfieChristl found the dirt: https://themarkup.org/...
  • @themarkup @themarkup on x
    In a rare look at just how deep (and weird) the rabbit hole of targeted advertising gets, we found 650,000 ways the digital ad industry labels you. And we've made it easy for you to explore the data: https://themarkup.org/...
  • @wolfiechristl Wolfie Christl on x
    A while back, I stumbled upon a file I consider the largest piece of evidence revealing how hundreds of data brokers trade personal data on everyone, including very sensitive data, globally. Massive investigation by @themarkup and German @netzpolitik_org: https://themarkup.org/..…
  • @themarkup @themarkup on x
    There are ways you can prevent companies from tracking you (insert https://themarkup.org/...) and avoid ending up in these segments in the first place. Or, if you are already being profiled and curious about your results, there are a few things you can do: https://themarkup.org/.…