When the Pipes Change, So Do the Rules: What OpenAds and AdCP Signal for Advertisers
By Lior Fisher | November 24, 2025
Programmatic’s plumbing is changing.
The codes and control points are moving closer to where decisions are made.
Lately, two moves stand out.
The Trade Desk’s publisher-side push with OpenAds and PubDesk, and the debut of Ad Context Protocol – AdCP – an open language for agent-to-agent coordination.
Different origins. Same implication.
The buying and selling of media is getting rewired around fewer middlemen, clearer signals, and machine-readable context.
Why this matters now
DSPs and SSPs grew up on scale, IDs, and speed.
That stack was built for auctions first, analysis second.
It optimized delivery. It didn’t always explain value.
Three gaps are becoming obvious:
1) Unclear supply paths – too many steps, too little clarity on who took what or why a bid won.
2) Shrinking signal quality – viewability and CTR tell part of the story, not the part that predicts outcomes.
3) Fragmented visibility – advertisers see their own delivery, not the competing environment their ads appeared in.
A lot of what we’re seeing now – OpenAds on the publisher side, AdCP on the coordination layer – are direct responses to those gaps. And they’re not happening in isolation.
The IAB Tech Lab recently released two new standards built for this shift. The first – the Agentic RTB Framework (ARTF) – is designed to make programmatic decisions faster and easier to audit by bringing more of the logic directly into the auction environment. The second – the User Context Protocol (UCP) explores how intent and context can be shared between systems without leaning on user IDs at all.
Different initiatives, same signal: the infrastructure is shifting toward cleaner paths, lower latency, and machine-readable context.
OpenAds and PubDesk – control moves up-stack
OpenAds is a fork of Prebid’s open-source auction wrapper, rebuilt by The Trade Desk PubDesk is the analytics layer that shows how The Trade Desk’s demand values inventory.
But there’s a trade(desk)off (sorry, had to insert that pun).
When a major DSP runs the wrapper, neutrality becomes a design question. And that’s a problem because transparency should be verifiable.
The takeaway for brands: directness works, but only if you can still benchmark. Closer pipes help, but you need cross-pipe outcome comparisons to optimize.
AdCP – a common language for agentic buying
AdCP sits above the auction.
It doesn’t replace OpenRTB’s fast bidding.
It standardizes the slow thinking around it – planning criteria, audience definitions, constraints, performance goals, and the reasoning behind decisions.
In practice, AdCP is an infrastructure for AI agents to coordinate without locking signals inside a single platform. Audit trails and human-in-the-loop controls are a default part of the spec.
That matters because agentic buying is coming, whether the industry is ready or not. When machines negotiate pacing, placement, and pricing, you will either see the logic – or you won’t.
AdCP is a healthy attempt to make that logic inspectable.
What to change if you’re an advertiser – yesterday
You don’t control who owns the wrapper or which protocol wins.
But you do control the signals you bring and the context you can read.
Shift three habits.
Report beyond delivery.
Keep the basics – reach, viewability, CPM.
Add the signals that predict outcomes:
– Time in view
– Refresh behavior
– Live user presence during exposure
– Placement quality and ad density per fold
– Competitive presence in the same moment
Those are the actual indicators of real attention.
Benchmark in both directions.
Within your buy: which contexts lift or drain attention.
Across the market: how your formats, placements, and creatives stack up against category peers.
Plan where your ads can actually be seen.
Adjust for agentic decisions.
If agents will plan and optimize, give them a structured context.
Name your objectives clearly. Expose the constraints you care about – viewable time, clutter thresholds, attention floors, geo performance bands.
Keep the audit trail – if a system changes pacing or placement, know why.
What to watch as this evolves
Neutrality tests. Any wrapper aligned with a major buyer must prove fairness in practice. Auditable logs and side-by-side auction diagnostics are not nice-to-haves.
Signal portability. If AdCP gains adoption, bring your attention and environment signals into that conversation. The winner will be the buyer with better inputs (no, not loudest spend)
Outcome alignment. Moving from impressions to attention will force new pricing models. Beyond inventory, be ready to buy time in view and clean context.
The quiet shift under our feet
Programmatic began by making supply liquid.
The next phase makes context visible.
OpenAds moves control closer to the source. AdCP makes decisions explainable across systems.
For brands, the advantage won’t come from picking sides in a wrapper war.
It will come from seeing more than your own campaign – the screen around it, the competitors beside it, and the conditions that make attention possible.
That’s the work.
Own your signals, widen your field of view, and benchmark what matters.
Latest Articles
-
The Doctor Everyone’s Seeing: YouTube, AI Search, and the New Rules of Visibility
Google’s AI Overviews are reshaping discovery by deciding which sources appear in answers — often before users ever click. As search shifts from clicks to citations, traditional performance metrics no longer tell the full story. This article explores what visibility really means in an AI-mediated web, and why competitive intelligence is now essential.
View Now -
The Vanishing Web: What Google’s Admission Really Means for Advertisers
For years, Google insisted the open web was thriving. Then, in a legal filing, it admitted the opposite. For advertisers, the real risk isn’t where ads run, but how much of what happens around them can still be seen, understood, and measured.
View Now -
What the RSL Standard Signals About the Future of Visibility in 2026
When Reddit, Yahoo, Medium, and several of the web’s biggest content platforms announced a new Really Simple Licensing (RSL) standard, most coverage focused on the politics: platforms finally demanding compensation from AI companies; a new legal framework for training data; the good old open-web fight. But if you take one step back, something bigger comes into focus. For the first time, publishers are trying to engineer visibility - not for users, but for AI agents. And advertisers should be paying attention.
View Now