In Episode 22 of The PubWay podcast, hosts Tina Iannacchino and Mike Villalobos welcome Brian Lin, SVP of Product Management, Advertising at TelevisaUnivision. The episode dives deep into the state of CTV measurement and the broader challenges facing programmatic advertising today.
Brian, who leads advanced advertising strategy for the world's leading Spanish-language media company, shares timely insights on first-party data, evolving consumer behavior, and what publishers and advertisers need to get right if they want to improve performance across CTV environments.
The conversation touches on everything from co-viewing dynamics and data match rates to the role of AI in scaling measurement. Below, we break down the most critical takeaways from the episode.
- CTV Measurement at a Crossroads.
- Closing the Gaps with First-Party Data.
- Measuring CTV in Multi-Viewer Environments.
- The Role of AI in Real-Time CTV Optimization.
- Looking Ahead: What’s Next for CTV Measurement?
- Realigning Expectations Around Performance.
CTV Measurement at a Crossroads
With connected TV approaching mass adoption, advertisers and publishers alike are trying to understand how to measure campaign performance accurately. One of Brian’s early points underscores the magnitude of this shift:
“CTV is almost at the point in which it's getting close to 50% of total video consumption.”
This growth brings both opportunity and complexity. While digital tools make CTV inherently more measurable than traditional linear television, many advertisers still struggle to capture the full picture of their campaigns. One reason? Data fragmentation.
Advertisers often rely on disparate data sets stitched together through intermediaries, introducing gaps and reducing match rates. “There’s always a tradeoff between data quality and scale,” Brian explains. “You want a high match rate, but not at the expense of accuracy.”
In CTV, where brands look to measure outcomes like cost per completed view (CPCV), return on ad spend (ROAS), and unique viewer reach, missing signals can undermine performance and accountability. Improving CTV measurement starts with improving the quality and interoperability of the data itself.
Closing the Gaps with First-Party Data
For publishers and advertisers, closing the measurement gap means building stronger, more privacy-conscious data infrastructure. Brian points to TelevisaUnivision’s own first-party data strategy as a blueprint.
By building a household graph that aggregates signals from across local live events, streaming content, linear television, and audio platforms, the company now reaches 95% of US Hispanics. “It’s a game changer,” Brian notes, especially in a landscape where third-party data still struggles to identify Spanish-speaking audiences accurately.
“Some third-party datasets show only about 40% accuracy in identifying Hispanic consumers,” he explains. “That’s a huge miss for advertisers with the right intent.”
Clean rooms are emerging as an effective solution to connect first-party data from publishers and advertisers. These environments allow datasets to be combined securely, enabling granular CTV measurement while respecting privacy standards.
Measuring CTV in Multi-Viewer Environments
Traditional measurement models were built for one-to-one devices like laptops and mobile phones. But with CTV, viewers gather in living rooms, often watching together. This creates a multiplier effect on impression value and a measurement blind spot for brands focused solely on device-level data.
“Co-viewing is still one of the biggest opportunities in CTV,” Brian says.
While general market co-viewing rates hover around 1.5 to 1.7 viewers per screen, that figure rises to 2.6 to 3 for US Hispanic households.
What this means in practical terms is that a CTV ad served to one device might actually be reaching three people. Adjusting measurement frameworks to account for co-viewing can dramatically improve perceived campaign performance, especially in family-oriented or multicultural households.
But to do so, publishers must be willing to share more metadata and log-level data with advertisers. “It’s a receipt,” Brian explains. “Advertisers should know what content their ads ran against if we want them to measure and come back.”
The Role of AI in Real-Time CTV Optimization
AI is already playing a supporting role in content classification, sentiment analysis, and targeting. But its true potential lies in making CTV measurement more dynamic and adaptive.
Take metadata, for instance. In the past, CTV inventory was often sold in bulk, with little transparency about the content it would appear alongside. But as AI tools improve, publishers can now categorize programming with greater precision, identifying not just genres, but tone, emotion, and thematic context.
This opens the door for more sophisticated brand safety controls and targeting strategies. For example, an advertiser promoting family products might want to align with upbeat, co-viewed programming but avoid more intense or adult-themed content.
At TelevisaUnivision, AI is also being applied in creative ways. During the Latin Grammys, the network partnered with ShopSense and Walmart to create a second-screen experience: as celebrities walked the red carpet, viewers could scan a QR code to shop similar outfits in real time. It’s a small but tangible example of how CTV advertising can evolve beyond traditional ad pods.
Looking Ahead: What’s Next for CTV Measurement?
As the podcast wraps, Brian offers a glimpse into a future shaped by both AI and, surprisingly, quantum computing.
With current cloud infrastructure, many platforms sample data rather than process it all, limiting the granularity of insights. But new breakthroughs in quantum hardware could allow real-time analysis of massive data sets without the tradeoffs publishers face today.
“Most programmatic partners don’t look at every opportunity in the bid stream because the cost is too high,” Brian explains. “With quantum computing, that could change.”
More immediately, publishers need to rethink how they define and share content metadata. While some hesitate to expose too much information for fear of cherry-picking, withholding it entirely limits advertisers’ ability to measure outcomes, target appropriately, and ensure brand safety.
The industry will likely move toward more transparency over time, driven by advertiser demand, technology improvements, and the increasing sophistication of AI tools that can enrich CTV metadata automatically.
Realigning Expectations Around Performance
With so many variables at play, CTV advertisers often ask a simple but important question: what’s the benchmark? Did my campaign deliver what it promised?
Today, many of those benchmarks are still being written. From completion rate to exposed audience to brand lift, CTV measurement still lacks the standardization of linear television. But progress is being made.
By embracing innovations like clean rooms, metadata enrichment, and cross-platform data graphs, publishers can offer advertisers the clarity they need. And when that happens, the entire CTV ecosystem becomes more efficient, accountable, and resilient.
As Brian puts it, “When advertisers get access to the right data, and can prove effectiveness, they come back.”
Tune In to the Full Episode
For a deeper dive into data quality, CTV campaign performance, and how publishers like TelevisaUnivision are shaping the future of digital video, listen to Episode 22 of The PubWay: Navigating Data Quality & CTV Measurement.
Share post
