By Filippo Gramigna, co-CEO, Onetag
Many have tried to change ad tech’s business model.
Not only to reduce fees or increase transparency, but also because, they argue, the current transactional system fails to align with business outcomes. The platforms that power the open web still operate like intermediaries, charging per impression or as a percentage of media spend, rather than delivering value as true infrastructure partners.
This volatile, margin-squeezing, difficult-to-forecast model makes investors nervous. Many of us have tried, and failed, to push the industry toward a more stable Software-as-a-Service (SaaS) structure, where cost and value are predictable, and incentives are clearer. I’ve experienced those attempts firsthand, both on the sell side and in conversations with brands and agencies. On one project I handled for a major laptop manufacturer, the agency was paid based on Full-Time Equivalents and aligned with KPIs such as conversion rates, rather than media spend. That worked well, as it aligned the interests of brand and agency.
Despite numerous examples like that, the industry as a whole hasn’t wanted to move. Might AI change the equation?
Why SaaS Never Took Hold in Ad Tech
The percentage-of-media model has stuck around for a reason. Advertisers like its flexibility — they can ramp spend up or down without committing to upfront costs. Agencies benefit, too, adjusting take rates or margins in response to performance or pressure. Everyone gets room to maneuver. No one has to commit to hard numbers.
It’s not that advertisers oppose transparency. But when Ari Paparo tried to shift Beeswax to a fixed-fee model, as he’s discussed publicly, their clients didn’t want it. They preferred the elasticity of media-based pricing — even if it made costs less predictable.
Meanwhile, agencies are in a bind. Many are now measured on business KPIs, but aren’t paid for hitting them. Their compensation is still tethered to media volume, not outcomes. And because they can’t guarantee how much a brand will spend in the future, they’re reluctant to adopt fixed-fee pricing with their own partners.
The current model benefits no one fully, but it works just enough to keep going.
What AI — and Agentic Systems — Could Unlock
That may not last. AI isn’t just automating tasks. It’s restructuring workflows.
As generative and agentic AI begin to influence planning, creative, and media execution, we may all move from fragmented, multistep processes to single-platform, multi-function solutions. That raises the question: do we still need different buy- and sell-side platforms? Or do we just need one intelligent entry point that connects brand objectives to curated media paths?
In a world of agentic media — where AI agents plan, create, and activate campaigns — the percentage-of-media model becomes harder to justify. The infrastructure, logic, and optimization are delivered as software. So why not price it that way?
Imagine a fixed monthly fee for always-on campaign orchestration, powered by AI and enriched with performance data. This would be infinite ad serving, with traffic shaping and outcome optimization baked in.
This isn’t just theoretical. Clinch recently launched a flat-fee subscription model that replaces CPM-based billing with unlimited ad serving. Coca-Cola and Hyundai are already testing it. It’s a small step — but it portends a deeper shift.
SaaS and the Efficiency Imperative
Curation is often seen as a media tactic. Done right, though, curation engineers efficiency: It filters the bidstream, reduces waste, and aligns supply paths to campaign KPIs — making automation more accurate and outcomes more predictable.
Yet most curated deals today are billed transactionally. They still live in a percentage-of-spend world. That’s fine for now — but as AI gets smarter and performance becomes more autonomous, those costs will start to look out of place.
Infrastructure should be priced like … infrastructure. Curation should be valued for what it delivers — not just how many impressions it moves.
What Needs to Change
To get there, we need to solve for three things:
- Shared incentives between brands, agencies, and platforms — tied to measurable business outcomes.
- Better forecasting of media needs and performance — which AI can help deliver.
- Transparent, infrastructure-level pricing that reflects the real cost of delivering clean, performant supply.
It won’t be easy. However inefficient it may be, the current model offers just enough flexibility to survive. But the more we automate, the more we’ll need to align economics with performance. And that may finally push ad tech toward the SaaS model investors have been hoping for.
Because in a world where AI agents do the work, the question won’t just be: How much media did you buy?
It’ll be: How efficiently did you deliver the outcome?
And that’s a question software is built to answer.
Originally Published on: LinkedIn
