In partnership with

You’ve heard the hype. Now it’s time for results

After two years of siloed experiments, proofs of concept that fail to scale, and disappointing ROI, most enterprises are stuck. AI isn't transforming their organizations — it’s adding complexity, friction, and frustration.

But Writer customers are seeing a positive impact across their companies. Our end-to-end approach is delivering adoption and ROI at scale. Now, we’re applying that same platform and technology to bring agentic AI to the enterprise.

This isn’t just another hype train that doesn’t deliver. The AI you were promised is finally here — and it’s going to change the way enterprises operate.

See real agentic workflows in action, hear success stories from our beta testers, and learn how to align your IT and business teams.

Our Amazing Sponsor

🚨 The Big Lie: "Anonymous user? No problem. We’ve got AI, machine learning, and unicorn dust."

Let’s not kid ourselves.

We built this entire digital advertising ecosystem on the promise of precision. Of automation. Of programmatic as the crown jewel of efficiency—a data-driven, machine-optimized way to reach the right person at the right time with the right message.

But here’s the plot twist:
Under the hood, a whole lot of it is just digital voodoo in a lab coat. 🧪🧙‍♂️

Vendors are out here claiming to turn anonymous bidstream crumbs—IP addresses, timestamps, maybe a lonely pageview on a sports blog—into robust, high-value audience profiles. And they’re not even breaking a sweat as they tell you this user is a: “Sports enthusiast, age 25–34, high-income, interested in Tesla, probably drinks organic cold brew, and votes in midterms.”

All that from what? A single click on Bleacher Report?
Come on. If your cousin shared a cat video on Facebook, would you assume he's an animal rights activist with disposable income? No? Then why are we okay with this logic when it's dressed up as audience science?

🤹‍♀️ The Alchemy Act: Turning Crumbs Into Clout

Here’s how the trick works:

  • Start with bidstream residue—things like a device ID, URL, maybe a geolocation ping from six weeks ago.

  • Throw it into a black box labeled “AI-powered segmentation” and hope no one asks questions.

  • Out pops a supposedly targetable persona, ready to be sold at a premium CPM.

Except—it’s not really a persona. It’s a fantasy.
The kind your middle school friend invented when he claimed he had a Canadian girlfriend you’d never meet.

It’s inference theater.
It’s programmatic cosplay.
And it’s completely antithetical to what programmatic was supposed to be.

We didn’t build algorithms and real-time bidding so someone could randomly label a user who glanced at a sports headline as a “luxury auto intender.”

That’s not targeting. That’s astrology with a data feed.

🔥 Why You Should Be Angry — Not Just Mildly Irritated

This isn’t just a technical issue—it’s a fundamental betrayal of trust.

  • 💸 Advertisers are flushing budgets down the DSP drain based on fabricated audience segments.

  • 👎 Consumers are being creepily stalked by ads that are both invasive and wildly off-target.

  • 🧼 The adtech industry is selling itself as a science, but under-delivering with smoke, mirrors, and inference spaghetti.

  • 🚽 Trust is circling the toilet, and if we’re not careful, what’s left of programmatic credibility will go with it.

And don’t even get me started on the “verified” vendors who double down on this nonsense while pretending their secret sauce is too complex to explain. If you can’t tell me how the sausage is made, I assume it’s mostly sawdust and old shoes.

Pesach Lattin, Publisher @ ADOTAT

🔍 How Ad Tech Vendors Recycle RTB Crumbs into Pseudo-Precision

Let’s get one thing straight: there’s nothing inherently magical about the bidstream.
No buried treasure, no mystical signals, no data goldmine waiting to be mined.

What does live in the bidstream? A bunch of technical metadata shared in the milliseconds it takes to serve an ad. Think: IP address, device type, page URL, user agent, timestamp, and sometimes location estimates. It’s the advertising equivalent of checking someone’s shoes and guessing their entire personality.

And yet, some vendors have turned this thin gruel into a marketing soufflé, promising advertisers they can build rich, hyper-targeted audiences out of practically nothing. Which is kind of like writing someone’s dating profile based on a parking ticket.

🧙 Alchemy 101: Spinning Bidstream Crumbs Into Personas

Here’s the playbook:
Take something small, cheap, and free-floating like an IP address or device model, and inflate it with just enough inference to sound scientific. Layer on third-party data stitched together from who-knows-where, toss it into a machine learning blender, and voilà—you’ve got yourself an audience segment.

Examples of this digital witchcraft include:

  • IP Address → Demographic Assumption:
    A user logs in from a Starbucks in Beverly Hills? Must be a high-net-worth Tesla fan with disposable income. Never mind that it could be a delivery driver checking the weather.

  • Page URL → Interest Category:
    Read one article about mental health? Congratulations, you’re now part of the “depression medication intent” cohort. Great news for the pharma brands—not so great for your privacy.

  • Device + Time of Day → Lifestyle Profiling:
    Browsing on a mobile device at 2 a.m.? Ad tech just labeled you an insomniac millennial with high anxiety and no impulse control. Here's an ad for sleep gummies and self-help ebooks.

It’s all smoke, mirrors, and a lot of "trust us, the model works."

🤦‍♂️ When Inferences Implode: The Real-World Fails

These aren’t just academic problems. Real campaigns based on inferred data go hilariously (and sometimes dangerously) off course:

  • 📉 Health Misdiagnosis by Proxy:
    A user clicks on a news article about cancer research, and suddenly they’re getting ads for sketchy “alternative” treatments. Interest is not identity, and this leap is both medically irresponsible and legally questionable.

  • 📍 Location-Based Stereotyping:
    Someone in a public library on the “wrong side of town” starts seeing a flood of payday loan ads. No one asked if they live there. No one cared.

  • 🏳️‍🌈 The "Gay Bulldozer" Incident:
    An audience labeled as LGBTQ+ got targeted with heavy equipment ads. Why? Because an adtech vendor's segmentation logic crossed a behavioral keyword stream and short-circuited. Intersectionality meets industrial supply chain.

The worst part? These failures don’t stop the money from flowing. Vendors just shrug. “Well, that’s what the model said.”

💸 The Fantasy Data Industrial Complex

An entire cottage industry thrives on these sketchy segments:

  • Data Brokers: Package and resell clusters like “Luxury SUV Dads” built off device IDs and vague intent signals.

  • SSPs/DSPs: Layer on targeting fees and segment markups to boost CPMs.

  • Audience Analytics Firms: Offer pixel-based “insight dashboards” that show you how your deodorant campaign reached 17% more “Urban Explorers.”

The value chain is built on what can only be described as digital fan fiction. And everyone makes money—as long as no one asks what’s real.

Verification? Not a chance.
Transparency? Only if you squint hard enough to read the footnote that says, “Data modeling subject to estimation.”

⏳ The Ethical Time Bombs

All this inference, profiling, and pseudo-targeting may seem like just another Tuesday in adtech—but it’s also a ticking compliance grenade.

  • 🧾 Consent Theater:
    Users didn’t consent to being tagged as depressed, broke, or politically volatile. Bidstream-based segments often skate around consent laws by pretending inferred data isn’t personal data. Good luck selling that to the regulators.

  • 🚫 Algorithmic Discrimination:
    When your ad tech labels someone as low-income and starts excluding them from credit or job ads? That’s not just shady—that’s legally actionable discrimination.

  • 🇪🇺 Regulatory Roulette:
    GDPR and CCPA require lawful basis for personal data use. Vendors lean into “legitimate interest” or claim the data isn’t personal. But when you can triangulate a person’s ZIP code, habits, and interests from a single auction ping? The legal gray zone starts looking awfully black and white.

  • 🧠 Consumer Distrust:
    When users start noticing eerily specific ads tied to content they barely skimmed, the backlash isn’t just about privacy—it’s about feeling surveilled. And that’s a brand killer.

🫅 The Emperor’s New Data

Let’s call it what it is: programmatic theater.
All the tools, terms, and tech dressed up to make advertisers believe they’re buying laser-precision audience access—when what they’re often buying is a wildly extrapolated guess based on minimal evidence.

What’s worse? Everyone involved knows this. They just keep playing along because it’s profitable and opaque. No one audits the segment logic. No one questions the persona math. We just write the checks, nod sagely, and pretend that “likely car buyer” isn’t just code for “clicked on an auto blog once six months ago.”

Until regulators step in—or marketers finally demand to see the receipts—we’ll be stuck in this loop, buying ad impressions based on aspirations, not reality.

The Legal Fiction Propping Up Programmatic Targeting

🧯The Mirage of "Informed Consent"

Let’s dispense with the polite fiction that people “consent” to their behavioral data being auctioned off in real time.

Yes, users click "accept" on a cookie banner. No, they don’t know that this means their browsing behavior, device metadata, geolocation, and more are being flung into a chaotic ad exchange that looks more like a Wall Street trading pit than a privacy-compliant system.

Real consent, under GDPR, must be freely given, specific, informed, and unambiguous.
And yet, most bidstream data—collected during real-time bidding (RTB) to inform ad auctions—is reused far beyond that initial transaction. It’s retained, enriched, modeled, and ultimately weaponized for profiling.

And under CCPA, where consumers must be allowed to opt out of the sale of personal data, most have no idea their data was "sold" in the first place. You can’t opt out of something you didn’t know existed.

So when ad tech vendors claim their audience segments are "consent-based"?
That’s a marketing line—not a legal one.

🌐 Borderless Data, Boundless Risk

Real-time bidding doesn’t care about national borders—but privacy law does. And here’s where things get radioactive.

Bidstream data, including GPS coordinates, page URLs, device types, and app usage, is frequently transmitted across international ad exchanges. Once it's out there, it can be accessed—or intercepted—by any party in the chain.

That’s where the Protecting Americans’ Data from Foreign Adversaries Act (PADFAA) comes in. It’s not just regulatory theater. It’s a reaction to very real fears that U.S. citizens' location data and behavioral metadata are ending up on foreign servers, potentially in China, Russia, or other adversarial nations.

And it’s not just theoretical. Investigations have shown that ad tech platforms have inadvertently exposed the movements of military personnel and government employees. This isn’t just a data ethics issue—it’s a national security liability.

You can bet regulators are paying attention. The question is: are you?

🔍 The Pseudonymization Myth

Let’s kill another comforting delusion while we’re here: the idea that “pseudonymized” means “anonymous.”

Ad tech vendors love to claim that their data is “safe” because it doesn’t include names, emails, or phone numbers. Instead, it’s tied to mobile ad IDs, IP addresses, or hashed identifiers.

But under European privacy law, pseudonymized data is still considered personal data—because it can be re-identified when combined with other data. Which is exactly what ad tech does all day long.

It’s not anonymous.
It’s just anonymish.

In fact, if a vendor or partner down the chain can re-identify a user—and in most cases they can—then you’re on the hook for treating that data as personal.

Saying “it’s only an ID” is like saying “it’s only a fingerprint”—technically true, but legally meaningless.

⚠️ Enforcement Is No Longer Theoretical

Until recently, the ad industry largely operated on the assumption that regulators wouldn’t understand what was happening—or care enough to act. That assumption is now dead.

  • The FTC has already cracked down on data brokers for misusing location data collected via bidstream, especially when it involved sensitive sites like abortion clinics or addiction treatment centers.

  • European regulators, including the Irish Data Protection Commission and the UK’s ICO, are sharpening their teeth on RTB’s lack of compliance. Multiple investigations are underway into how consent is being faked or bypassed.

  • Class action lawsuits are gaining steam, targeting companies that used bidstream data to profile consumers without consent. Some privacy advocates are calling this the largest data breach in history—not a single incident, but a structural failure across the ad tech ecosystem.

If you think this won’t end in high-profile legal bloodshed, you haven’t been paying attention.

🧨 The Hidden Liability for Advertisers

Here's the part that should make your legal team sweat:
Advertisers are not innocent bystanders in this mess.

When you buy segments labeled:

  • “Chronic Pain Sufferers”

  • “Financially Distressed Parents”

  • “Discreet LGBTQ+ Travelers”

…you’re entering a legal contract based on profiling. And if that profiling was built on improperly obtained data, you’re co-liable.

Regulators don’t care that you didn’t build the segment yourself. If your campaign targeted a protected class using inferred personal data, and you can’t prove lawful basis or consent? You’re in the crosshairs.

It’s the digital equivalent of buying a Rolex out of the trunk of a car and acting surprised when the police show up.

🧭 A New Compass for Targeting in a Legal Minefield

So what now? Are we just doomed to keep building audience segments on a foundation of quicksand?

Not necessarily. There are alternatives—ethical, transparent, and legally sound:

  • Contextual Targeting:
    Ads based on page content don’t require behavioral data and still drive relevance.

  • First-Party Data:
    Consent-driven, direct relationships with users are harder to scale, but infinitely more trustworthy.

  • Consent Management Platforms (CMPs):
    If you must use personal data, use tools that give users real choices and retain proof of consent.

  • Vendor Auditing:
    Ask the hard questions: Where did this segment come from? Is it modeled or observed? Was consent collected? Show me the chain of custody.

Because at this point, turning a blind eye isn’t just irresponsible—it’s dangerous.

🧠 Final Thought: Privacy Is a Legal Right, Not a UX Option

Too much of ad tech still treats privacy as an inconvenience to work around, rather than a principle to be protected. But consent isn’t a hurdle—it’s a contract. And if that contract is built on deception, the entire system will collapse under legal pressure.

The era of plausible deniability is over.
The age of accountable targeting is here—whether we like it or not.

🚨 Why Upgrade to ADOTAT+?

Because telling your boss you missed this week’s biggest adtech story is a lot more expensive than a subscription.

Because not knowing that 40% of your campaign “audience” might be bots is a great way to get promoted… to consumer support.

Because if you don’t read it, someone else at your agency will—and then take credit for spotting that your “luxury auto intenders” were actually just interns researching for a blog.

Because explaining CPM inefficiencies, log-level data demands, and why “Yoga Moms Who Own an SUV” aren’t a real thing… requires more than vibes.

ADOTAT+: It's like a lie detector test for the entire digital marketing ecosystem.

logo

Subscribe to our premium content at ADOTAT+ to read the rest.

Become a paying subscriber to get access to this post and other subscriber-only content.

Upgrade

Keep Reading