A new joint study from the Coalition for Innovative Media Measurement and the 4As surveyed 197 senior marketers and conducted 16 executive interviews to diagnose the state of audience measurement across TV and video. The headline finding is that advertiser confidence hasn't kept pace with the sophistication of the tools now available. Forty-three percent of respondents rated cross-platform measurement and methodology transparency as major or severe barriers to effective measurement. Eighty-four percent identified AI as the most consequential technology on the horizon. The study calls for shared definitions, methodology transparency, and interoperable infrastructure as the path forward.
The gap between ad spend and ad tech is already producing waste
The measurement problem isn't theoretical, It's actively destroying capital.
A Samba TV report published in February found that for 12 of the top 20 TV advertisers, increased investment actually produced fewer unique household impressions than the prior year. More money in, less impact out. The report traced the disconnect to a fundamental mismatch between the audiences targeted by TV campaigns and the consumers actually in-market for those products. In the car insurance vertical, TV ads were being served against aspirational lifestyle audiences interested in fitness and travel while the households actively searching for policies online skewed toward domestic interests like parenting and home décor. The targeting infrastructure told advertisers one thing. The purchase behavior data told a different story.
This is what the CIMM study's "confidence gap" looks like when it hits a P&L. Advertisers aren't just struggling to reconcile different measurement systems. They're making allocation decisions inside a fog, and the fog is expensive.
A separate Advertiser Perceptions survey reinforces the scale of what's at stake: 70% of streaming advertisers plan to increase budgets by an average of 17% this year (2026), reallocating from linear TV, search, and social. That's core budget moving into an ecosystem that the same advertisers describe as a fractured mess. An earlier Mediaocean report from January puts a finer point on the execution gap: while 86% of marketers say cross-channel orchestration is a priority, only 10% have the unified systems to do it. Ultimately, this means that the money is moving faster than the plumbing.
Viewers are already telling you the ad experience is broken
FreeWheel's Voice of the Viewer study surveyed 4,000 European viewers and found that 72% say ad relevance hasn't improved over the past five years. That timespan covers an entire generation of programmatic infrastructure investment, billions in ad tech buildout, and the rise of machine learning-driven audience segmentation. The viewer's lived experience of receiving a relevant ad flatlined through all of it.
More than half of streaming viewers report seeing the same ads repeatedly in a single session. Advertising Video on Demand (AVOD) audiences are 10% more likely to find ads intrusive compared to linear TV viewers, with 39% of those who feel disrupted blaming (b)ad timing — interrupting dialogue or breaking narrative flow.
Compare the FreeWheel data against what's happening inside the 2026 FIFA World Cup: Telemundo sold 90% of its Spanish-language inventory more than six months before kickoff, doubling ad spend from 2022, with nearly 60 advertisers committed. Eight in ten Spanish-language viewers say World Cup ads are an enjoyable part of the experience. Same medium, same ad-supported model, radically different viewer response. The difference is that the World Cup creates native alignment between content, audience, and commercial intent. The advertising works because the context becomes the targeting that programmatic infrastructure has struggled to deliver.
Nearly half of the viewers in FreeWheel's study deliberately chose an ad-supported plan. Think of that tolerance as a checking account and you'll start to see there are insufficient funds. Streaming apps who are investing in the ad experience will keep the account in good standing with their audience and they'll command premiums for it. The ones relying purely on segmentation and frequency management inside contextually mismatched inventory will watch satisfied viewers quietly upgrade to ad-free tiers, and the CPM economics will stop working for everyone downstream (tricky business).
As industry advisor and founder of Harrison X Ventures Daniel Harrison frames it 'there is today revenue and tomorrow revenue'.
Daniel emphasizes finding the balance between today revenue (tactical, get the deal across the finish line) and tomorrow revenue (strategic growth opportunities) and this is that dynamic playing out in streaming.
AI is the most anticipated solution to a problem it can't structurally solve
The CIMM study's most striking data point is that 84% of advertisers see AI as the most consequential technology development for measurement. More than any other advancement on the horizon.
But there's a category error embedded in that optimism. AI excels at pattern recognition within a dataset. It doesn't necessarily do a great job of resolving upstream disagreements about what datasets should measure, how identity should resolve across platforms, or which methodology governs when two systems produce conflicting outputs. The CIMM study itself identifies the core problem as competing "versions of the truth" across incompatible systems. AI can move faster inside any one version of the truth but it can't decide which version is right.
Meanwhile, the platforms best positioned to deploy AI against measurement aren't using it for that purpose. FreeWheel's study found that more than half of viewers are open to AI playing a role in ad selection. YouTube is testing conversational AI tools on smart TVs. Amazon is pushing Alexa+ onto Fire TV with expanded interaction capabilities. Roku and Netflix are both investing in AI-powered discovery. The AI investment is flowing into content engagement and retention instead of the measurement reconciliation that advertisers say they need most.
The industry wants alignment. The market is rewarding the opposite.
The CIMM study recommends shared definitions, methodology transparency, and interoperable infrastructure. Those are the right solutions for a cooperative ecosystem. The problem is that the companies with the most complete data infrastructure have no incentive to cooperate.
Amazon's ad business grew 22% in Q4 2025, reaching $21.3 billion. That growth was built on exactly the kind of closed-loop system the CIMM study implicitly argues against: proprietary purchase behavior data, exclusive access to Prime Video and NFL Thursday Night Football inventory, and a DSP that reportedly discounted fees to as low as 1% for major agency spenders. Amazon doesn't need cross-platform interoperability because it owns enough of the stack — content, commerce data, targeting infrastructure, attribution — to operate as a self-contained advertising economy.
Jeff Green, CEO of The Trade Desk believes this could lead to antritrust challenges for Amazon and a future without an Amazon DSP.
Netflix is building the same architecture from a different entry point. In the span of 44 days this year, the company opened its ad inventory to audience targeting via the Amazon DSP and Yahoo DSP, launched a proprietary Conversion API with early attribution results that outperformed benchmarks by more than 75%, consolidated product, engineering, and data under a single executive, and acquired an AI production technology company. Netflix now controls content creation, audience relationship, targeting data partnerships, measurement, and inventory surfaces. It doesn't need to declare itself an ad tech company. It just needs to keep building a system where interoperability with the rest of the ecosystem is optional.
The Trade Desk represents the most ambitious attempt to build a unified intelligence layer for the open internet. Its OpenTTD portal consolidates every partner function into a single analytics layer. But the market has rendered its verdict as growth decelerated from 25% to 10%, the stock shed roughly 80% from its all-time high, and the company announced a unified intelligence portal in the same week it told Wall Street its best growth days may be behind it. Nielsen's push into 200+ advanced audience segments on Nielsen ONE is a play for relevance in the same direction — behavioral targeting, outcomes over eyeballs — but it's still a vendor selling into an ecosystem where the largest buyers are building their own.
The structural reality is that fragmentation isn't a bug the industry is trying to fix. For Amazon and Netflix, it's a feature. Their competitive advantage is the intelligence layer they control, and every dollar an advertiser spends trying to reconcile competing measurement systems across the open internet is a dollar that could have been spent inside a walled garden where attribution is native and reconciliation is unnecessary. The CIMM study asks the industry to build bridges. The market is rewarding companies that build walls.
That's the tension the measurement conversation keeps dancing around. The 43% of advertisers who rated cross-platform measurement as a severe barrier aren't going to solve that problem through industry working groups and shared definitions — not while the companies with the most complete data sets are actively consolidating the intelligence infrastructure around themselves. The measurement gap will start to close when someone builds a system comprehensive enough to make reconciliation unnecessary. Amazon already has. Netflix is getting there. Everyone else is writing white papers about interoperability while the window to build an alternative narrows.
This Article By the Numbers
- 197 senior marketers surveyed in the CIMM/4As study
- 43% of advertisers rated cross-platform measurement as a major or severe barrier
- 84% identified AI as the most consequential measurement technology
- 12 of 20 top TV advertisers spent more and reached fewer households (Samba TV)
- 72% of viewers say ad relevance hasn't improved or has worsened in five years (FreeWheel)
- 70% of streaming advertisers plan to boost budgets by 17% in 2026 (Advertiser Perceptions)
- 10% of marketers have the unified systems to execute cross-channel orchestration (Mediaocean)
- 22% Q4 2025 ad revenue growth for Amazon, reaching $21.3 billion
- 75%+ — how much Netflix's Conversion API outperformed attribution benchmarks
- 80% decline in Trade Desk stock from its December 2024 all-time high
- 8 in 10 Spanish-language viewers say World Cup ads are an enjoyable part of the experience
- 90% of Telemundo's World Cup Spanish-language inventory sold six months before kickoff
And if you enjoyed this article, you may enjoy one of these, too:
- Is Netflix Building an Ad Tech Company? It Sure Seems So.
- Blood Red Moon: The Trade Desk Built a Platform, FPC Wants to Build a Way Around It, and Amazon Is Still Winning.
- FreeWheel Study: Viewers Accept Ads, Experience Falls Short
- TV Ads Are Missing the Mark by Billions, New Report Finds
- Advertisers Flood CTV With Cash, But Demand a Better Return
- Marketers Bet Big on AI and CTV, But Their Tech Is Lagging
- The 2026 World Cup's Real Story Is Who's Building Infrastructure for the Next One
