Bauer Media Group shut down its German digital publishing operation and treated AI Overview traffic displacement as permanent, not cyclical. The numbers tell the rest.
PayPal launched an advertising identifier built on 26.3 billion verified transactions. The company that kept your purchases private now sells what you bought.
Identification Threshold Theory suggests the industry has been measuring exposure frequency while ignoring the only variable that actually determines whether an ad registers: whether each exposure lasted long enough for the brain to identify it.
The FTC's Mobilewalla complaint reveals how consent theater operates at scale — vague contractual assurances, three-app audits, and bid stream harvesting as a business model.
The document that precedes every campaign is broken — 33% of budgets wasted, the committee effect, and why the industry's strategic foundation is rotting from the inside.
Analysis of how LinkedIn became the second most-cited source by AI models for professional queries, with 89,000 distinct URLs cited across 325,000 AI prompts.
Misinformation publishers copy authorized seller IDs into their ads.txt files, laundering their inventory through the same programmatic pipes that brands pay for. The research shows the problem is structural, not technical.
Eye-tracking research from Simonov, Valletti & Veiga finds hard news produces equal or greater advertising effectiveness than soft news — challenging the $2.8B brand safety avoidance industry built on untested assumptions.
Meta surpasses Google in global ad revenue for the first time. The $243.46B forecast is not a victory — it's a confession about what the industry has become.
TikTok complies with the DSA's minor profiling prohibition. It also extensively profiles minors. These two facts are not in conflict — they describe different things.
Tencent patented a system that infers minor status from behavioral signals alone — cursor speed, social graph, session timing — without explicit input. The same infrastructure that protects also profiles.
The widening delta between content that is efficiently produced and content perceived as genuinely made — a consumer recalibration the ad industry cannot yet measure.
Amazon filed an IPR challenging the Datonics behavioral targeting patent while simultaneously operating its own ad network on the same mechanism. The legal and structural contradictions at the center of the profile-distribution targeting model.
Meta's March 2026 reclassification of engage-through conversions reveals the structural gap between platform-reported and true ROAS — and exposes how cross-platform benchmarking has always been comparing different rulers.
MIT's formal proof that AI advertising systems tend toward exploitation of the gap between perceived and actual product quality — a mathematical framework for understanding when personalization becomes manipulation.
Privacy Sandbox delivered neither the targeting value users tolerated nor the protection they were promised. Then Google closed the bridge before anyone crossed.
Adlook's September 2024 study found 82% of users targeted as 'Women 18-24' were not. The ad targeting system's demographic accuracy problem is structural, not fixable with better data.
YouTube's May 12 update will auto-insert mid-rolls into videos uploaded before February 24 — treating evergreen content as perpetually renewable commercial inventory.
Perla et al. found psychological targeting has near-zero end-to-end effectiveness. The ad industry pays a continuous tax to operate a machine that doesn't work.
How Google removed Display and Video campaign support from Performance Planner — and what it reveals about the infrastructure of awareness advertising inside a company that no longer needs awareness to function.
How the accounting system that measures only performance marketing declared brand equity invisible — and what the resulting budget allocation actually costs.
Platform-reported ROAS runs 30-50% higher than true ROAS. The fashion brand had a dashboard showing 5.0x. Their bank account said 0.95x. Nobody was auditing the gap.
The FTC banned Gravy Analytics from selling sensitive location data. The RTB infrastructure that delivered it keeps running. This is what enforcement looks like when it targets a company instead of a practice.
When Paladone got 88 clicks from Amazon's AI shopping assistant versus 500,000 from traditional ads, the CPC paradox revealed something structural about what advertising is becoming.
Apple Maps launched the first ad platform that targets situational context — search query, approximate location, on-screen view — rather than user identity. What happens when moments replace profiles as the primary targeting unit?
The FTC named what RTB always was: a systematic broadcast infrastructure for consumer data. The Mobilewalla case is the first time the legal system looked directly at the architecture.
The average K-12 student uses 67 EdTech applications per year inside a consent pipeline that was structurally broken before the first app was installed. Utah AG data meets developmental psychology.
Amazon's share of US retail media budgets dropped ten points in one year — from 56% to 46% — while the overall market grew. The money didn't disappear. It redistributed.
Brand safety blocklists are eating premium publisher inventory — the tools deployed to protect brands are actively defunding the journalism they depend on.
YouTube's pivot to 90-second unskippable ads on CTV isn't an experiment — it's the resurrection of television's interruption infrastructure, ported into an environment that once promised something different.
While regulators debated COPPA consent dialogs, AT&T was granted patent US10820051B2 for a system that uses biometric sensors to estimate a viewer's age and automatically swap ad content in real time.
Magnum pumped chocolate aroma into a London Tube tunnel in March 2026. Commuters described the result as urine mixed with chocolate. This is what happens when ambient advertising escapes the controlled conditions of a brief.