Experimental ad budgets are being restructured across the industry as AI platforms introduce new ad products and traditional performance channels h...
The Signal
Experimental ad budgets are being restructured across the industry as AI platforms introduce new ad products and traditional performance channels hit saturation. Marketers are reallocating test-and-learn dollars toward generative search optimization, ChatGPT pilot ads, retail media, and out-of-home — a meaningful departure from the performance-first orthodoxy that has governed digital spend for the past decade. The budget math is shifting in concrete terms. Agencies like Markacy, historically operating on an 80-20 split between core media and experimental spend, are actively revisiting that ratio. The pressure to find new audiences is driving willingness to accept looser KPIs and longer attribution windows in exchange for scale and incrementality. In some cases, experimental budgets are growing in absolute terms; in others, dollars are being pulled directly from channels previously considered guaranteed performance vehicles. Underpinning the reallocation is a broader infrastructure shift. NLP pipeline maturation and advances in video understanding models — machines now capable of interpreting motion, actions, and events across time — are expanding the inventory and targeting capabilities available through AI-native ad platforms, giving budget experimentation a more credible technical foundation than existed even twelve months ago.
Why It Matters
The 80-20 rule wasn't a strategy — it was a comfort blanket, and the industry is finally admitting it. For years, that ratio survived not because it was optimal but because performance channels delivered clean attribution numbers that made CFOs comfortable. The real story here isn't that experimental budgets are growing; it's that the definition of "guaranteed performance" has become indefensible as CPMs on Meta and paid search have inflated past the point where the math closes for most advertisers. What surface coverage misses is the infrastructure argument. The shift toward AI-native ad products and retail media isn't just opportunistic budget wandering — it's being pulled forward by genuine technical capability. Video understanding models that can interpret motion and context across time, combined with maturing NLP pipelines, mean the targeting and inventory quality on these emerging platforms is structurally different from what it was even four quarters ago. Marketers aren't just accepting looser KPIs out of desperation; they're making a credible bet that attribution will catch up to the inventory. The competitive dynamic worth watching is what happens to agencies still defending legacy channel allocations to clients. Markacy revisiting its 80-20 split is a signal, not an outlier. Shops that can demonstrate incrementality frameworks sophisticated enough to justify experimental reallocation will retain and expand mandates; those still selling guaranteed-performance narratives on saturated channels are defending a position that erodes every quarter. For decision-makers, the window to build institutional fluency in GEO, retail media, and AI ad products is right now — before these channels move from experimental line items to core budget obligations and the learning-curve advantage disappears.
The Exploit
Opportunity
Marketing leaders at mid-market and enterprise brands who build incrementality measurement frameworks now — before Q3 2026 — will capture a structural advantage: the ability to justify experimental reallocation to CFOs with credible attribution logic rather than gut instinct. Teams with that infrastructure can move budget decisively into GEO, retail media, and AI ad products while competitors are still seeking internal approval.
Risk
Agencies and in-house teams still selling Meta and paid search on guaranteed-performance narratives are defending a position that inflates in cost and erodes in credibility every quarter. The failure mode is losing mandate expansion to competitors who can demonstrate incrementality — not because the channels stopped working entirely, but because they can no longer justify the allocation math at current CPMs.
The Move
Commission an incrementality audit of your current channel mix before June 2026, with explicit authority to reallocate up to 30% of core media spend based on findings. The CMO owns this, not the media director. Success criteria is a defensible attribution framework that can withstand CFO scrutiny — one that makes experimental channels comparable to legacy performance channels on the same measurement terms.
First-Mover Advantage
Gains
Risks
Agencies and in-house teams still selling Meta and paid search on guaranteed-performance narratives are defending a position that inflates in cost and erodes in credibility every quarter. The failure mode is losing mandate expansion to competitors who can demonstrate incrementality — not because the channels stopped working entirely, but because they can no longer justify the allocation math at current CPMs.
Window
Winners & Losers
Winners↑
Retail Media Networks
As experimental dollars migrate away from saturated performance channels, retail media sits at the intersection of credible first-party data and incrementality storytelling — exactly what budget-anxious CMOs need to justify reallocation to CFOs. The channel benefits from a dual tailwind: it absorbs dollars fleeing Meta and paid search while simultaneously gaining legitimacy as AI-native targeting matures around it. Retail media operators should accelerate self-serve measurement tooling and cleaner attribution reporting to convert experimental line items into core budget obligations before Q3 2026 planning cycles lock in.
AI-Native Ad Platforms (ChatGPT Ads, Generative Search)
These platforms are the direct destination for the reallocation wave described here, and the technical credibility argument — maturing NLP pipelines and video understanding models that can interpret motion and context across time — gives them a stronger pitch than experimental channels have historically enjoyed. Advertisers are not just parking dollars speculatively; they are making a structural bet that attribution infrastructure will catch up to inventory quality. The strategic priority for these platforms in Q2 2026 is publishing robust incrementality case studies fast enough to accelerate the transition from pilot line items to committed spend.
Incrementality Measurement and Attribution Firms
The entire budget reallocation thesis depends on a credibility gap closing — marketers are accepting looser KPIs today on the explicit expectation that attribution methodology will mature to validate their bets. Firms that can deliver incrementality frameworks sophisticated enough to bridge experimental and core budget conversations are the connective tissue of this entire shift, and their value proposition has never been more directly aligned with a C-suite priority. These vendors should be packaging turnkey incrementality solutions specifically designed for GEO, retail media, and AI ad products and positioning them as the justification layer agencies need to defend reallocation decisions to clients.
Forward-Leaning Full-Service Media Agencies
Agencies that move now to build institutional fluency in GEO, retail media, and AI ad products will hold a durable learning-curve advantage as these channels transition from experimental to core budget obligations — the window is open in Q2 2026 but closes quickly once the category matures. The competitive dynamic is winner-take-most: mandates will consolidate around shops that can demonstrate credible incrementality frameworks, while laggards defending legacy channel allocations will face client attrition on an accelerating timeline. The immediate action is investing in dedicated practice areas for AI-native ad products and producing client-facing ROI narratives that reframe experimental spend as infrastructure investment rather than risk.
Losers↓
Legacy Performance Marketing Agencies
Agencies still anchored to guaranteed-performance narratives on Meta and paid search are defending a position that erodes structurally every quarter as CPMs inflate and clean attribution numbers become harder to produce at acceptable ROI. The Markacy 80-20 example is a leading indicator, not an outlier — shops that cannot build and communicate credible incrementality frameworks for emerging channels will lose both the mandate expansion conversations and, eventually, the core retainer. The defensive play is to acquire or partner with measurement specialists immediately and reframe client reporting around incrementality rather than last-click attribution before Q2 2026 budget reviews force the conversation.
Meta and Paid Search Channels
These channels are the explicit source of the dollar outflow described here — CPM inflation has pushed the performance math past the point of closure for a growing share of advertisers, and the 'guaranteed performance' positioning that justified dominant budget allocation is losing its hold as CFOs demand better unit economics. The mechanism of pressure is self-reinforcing: as experimental dollars leave and prove incrementality elsewhere, the justification for premium CPMs on saturated inventory weakens further. Both platforms will need to accelerate AI-native product development and offer more credible incrementality tooling of their own to slow the reallocation trend before it becomes a structural budget shift rather than a cyclical one.
Strategic Outlook
The timing matters because the technical preconditions for this shift finally exist. For years, experimental budget arguments were essentially faith-based — marketers accepting weaker attribution in exchange for hypothetical future scale. What's different entering Q2 2026 is that the infrastructure layer has caught up enough to make the bet credible. NLP pipeline maturation and the emergence of video understanding models capable of interpreting motion and context across time aren't incremental improvements; they represent a qualitative change in what AI-native inventory can actually deliver on targeting precision. Experimental channels now have a legitimate technical foundation, not just a growth narrative. Historically, budget experimentation spikes during two conditions: when incumbent channels get expensive, and when new inventory hits minimum viability. Both conditions are present simultaneously right now, which is unusual. The last comparable moment was the programmatic shift circa 2013-2015, when display inventory quality crossed a threshold that justified structural reallocation away from direct buys. The connection to retail media is underappreciated. Retail media's rise normalized the idea that closed-loop attribution could exist outside Google and Meta's walls — it made CFOs conceptually comfortable with alternative measurement frameworks. That precedent is now being extended to GEO and AI ad products, which is why the KPI flexibility conversation is happening faster than it otherwise would. Agencies defending legacy allocation ratios aren't just losing a tactical argument; they're losing the measurement credibility argument too.