Matt Wasserlauf, CEO, Blockboard
For years, the advertising industry had a convenient villain: fraud. It was real, it was measurable and it gave everyone something to blame when performance fell short of expectations. Brands hired verification vendors, agencies added line items to media plans and the industry collectively agreed that the problem was bots, not the system.
Fraud is still real. But as AI-driven automation removes layer after layer of manual execution from media buying, a different problem is coming into focus. Something harder to name, harder to fix and, for a lot of people in the supply chain, much more uncomfortable to talk about: structural waste.
The question isn’t whether the system is inefficient; it’s whether the industry is ready to see exactly how inefficient it is.
The programmatic system is working as designed
The modern programmatic stack was built incrementally, by different stakeholders, each with their own economic interests. DSPs, SSPs, data providers and verification vendors each take a margin. By the time a dollar enters the system and reaches a real human being watching a real screen, a significant portion of that dollar has been absorbed by infrastructure that the advertiser never fully sees and rarely fully understands.
This isn’t a secret. The industry has known about the ad tax (the cumulative take rate of intermediaries) for years. But knowing about something and being forced to confront it are different things. For most of the programmatic era, the math was fuzzy enough, and the market was growing fast enough that the inefficiency could be tolerated.
That era is ending.
AI is a transparency forcing function
When humans were managing campaigns manually — setting bids, reviewing placements, making optimization decisions — there was natural friction in the system. That friction created ambiguity, and ambiguity created cover. Nobody expected perfect visibility into every dollar.
AI doesn’t work that way. When machines are making thousands of buying decisions per second, the economics have to be precise. Automation surfaces everything: the fees that were buried in IO terms, the margin that was folded into CPM pricing, the intermediaries that added latency without adding value. Marketers can’t optimize around what they can’t see, so AI makes everything visible.
And visible math is very uncomfortable for a lot of people.
AI doesn’t just optimize performance; it audits the system it operates inside — that’s the part nobody is talking about at Cannes this year.
What the holding company tensions are actually about
The recent friction between major holding companies and DSPs isn’t primarily about pricing; it’s about visibility. Buyers are starting to demand transparency into the economics of the platforms they use — and platforms that built their margins on opacity are not happy about it.
The public positioning from both sides has been predictably diplomatic. But underneath the press releases, what’s happening is straightforward: AI-enabled buyers now have the data to ask questions that weren’t practical to ask before. How much of marketing spend is reaching inventory? What’s the effective take rate? Where is the money going between my order and the impression?
These are not radical questions. They are the questions every advertiser should have been asking for years. The difference now is that the answers are available, and the industry will have to decide what to do with them.
The conversation happening in the side rooms at Cannes
At Cannes this year, the most valuable conversation isn’t happening on any main stage. It’s the one happening over dinner, between senior brand-side buyers and the partners they actually trust, about what they’re seeing in their own data.
The questions are getting sharper. The tolerance for vague answers is getting shorter. And the technology now exists to give buyers real visibility — not modeled, not estimated, but verified — into where their dollars go and what those dollars actually accomplish.
For instance, Blockboard built its platform on the assumption that buyers deserve to see everything. No invalid traffic before a bid is placed. No hidden fees. No margin folded into delivery. The model is simple because it has to be: if a marketer can see everything, they can trust everything. And if they can’t see it, they should ask why not.
What comes next
The industry is at an inflection point. AI has given buyers the tools to demand real transparency. Some platforms will embrace it. Others will resist it for as long as they can. But the direction is clear, and Cannes, this year more than most, is a good place to have the honest version of this conversation.
Where did the money go? It’s not a rhetorical question anymore.
Partner insights from Blockboard
More from Digiday
How a German publisher JV is turning LLM visibility into a premium brand buy
Germany’s BCN, the joint-venture commercial arm of three major publishing houses – Hubert Burda Media, Funke and Klambt – is rolling out a commercial product that helps brands get properly surfaced and described inside ChatGPT, Gemini and other AI assistants, not just on traditional search results pages.
AI podcast experiments march on with Forbes’ new daily audio briefing
Forbes bets on AI-generated audio with a five-minute daily news brief. Stories are selected by product, editorial and an internal AI tool.
How USA Today Co. is trying to beat AI Overviews on World Cup news
USA Today Co. is using AI tools to beat AI Overviews in the race for World Cup search traffic around breaking news.