App Science Audiences Deliver Higher Attention Than CTV Averages, Study Finds

As connected TV shifts from a reach-first channel to a performance-driven medium, attention … not impressions … is emerging as the metric advertisers care about most. New research from TVision suggests that CTV effectiveness is increasingly shaped by audience strategy, not just premium inventory labels. The study found that campaigns using App Science audience segments delivered higher attention and co-viewing rates than CTV industry averages, signaling a broader shift in how brands, agencies, and platforms measure real engagement as CTV buying accelerates into 2026.

According to multiple TVision studies released this week, campaigns executed by Sabio using App Science® audience segments delivered higher levels of viewer attention than CTV industry benchmarks over the same period. The findings point to a growing opportunity for advertisers to move beyond generalized targeting and toward data-driven audience alignment that maximizes engagement, not just impressions.

The research focused on campaigns for a telecommunications advertiser that leveraged several of App Science’s proprietary audience segments, including Hispanic adults, tech enthusiasts, Baby Boomers, and Android users. These segments were paired with Spanish-language creative, an alignment that proved central to performance gains.

Across the studies, campaigns using App Science audiences exceeded CTV averages for both attention and attention-to-duration metrics. In some cases, inventory placed by Sabio delivered attention-to-duration ratios as much as three percent higher than the broader CTV average—an incremental gain that can translate into meaningful performance lift at scale.

Why Attention Is Becoming a Battleground Metric

For years, CTV has been positioned as a premium environment by virtue of its association with television-like viewing. But as the ecosystem fragments across hundreds of apps and formats, advertisers are increasingly questioning whether “premium” labels actually correlate with audience engagement.

TVision’s findings complicate that assumption. The research showed that Sabio-placed inventory was more likely to be viewed in co-viewing environments than inventory categorized as premium, including hybrid SVOD/AVOD apps and AVOD-only apps. Co-viewing rates exceeded hybrid SVOD/AVOD benchmarks by four percent and AVOD benchmarks by eight percent.

That matters because co-viewing—multiple people present in the room during ad exposure—can significantly amplify the real-world impact of a single impression, particularly for categories like telecommunications, retail, and entertainment.

“Attention and co-viewing are critical metrics for advertisers seeking to maximize CTV effectiveness,” said Hassan Babajane, chief revenue officer at TVision. “When viewers are present, engaged, and watching together, advertising performs better. Inventory that consistently delivers those conditions creates measurable value.”

Audience Strategy Over Inventory Labels

One of the most notable implications of the research is that audience targeting, when paired with relevant creative, may outperform traditional assumptions about inventory quality. Rather than relying on app-level prestige or content genre alone, the campaigns focused on aligning message, language, and audience context.

Helen Lum, EVP of App Science, said the findings reflect a broader shift in how CTV performance is being evaluated. “As CTV continues to mature, attention has become a key indicator of true campaign effectiveness,” Lum told Street Fight. “This research validates the impact that strategic audience segmentation can have on campaign performance.”

She added that App Science’s approach is grounded in real-world engagement behavior rather than surface-level targeting signals. “App Science leverages data-powered audiences uniquely built around how people are actually engaging with content, allowing advertisers to drive stronger attention and more meaningful impact.”

App Science operates as an independent data entity within Sabio, powering audience segments that are exclusive to the Sabio platform. Its data foundation combines mobile device and CTV streaming signals across more than one billion device IDs, anchored by a household graph representing roughly 80 million U.S. households—about 70 percent of all streaming households.

For agencies and brands, that scale is increasingly important as advertisers seek audiences that reflect the full diversity of the U.S. population, not just subsets overrepresented in certain apps or content environments.

Measurement Moves to the Forefront

The research also underscores how measurement expectations in CTV are changing. TVision’s methodology captures person-level data, including presence in the room, co-viewing, and second-by-second attention—metrics that go well beyond traditional completion rates or impressions.

As more brands push for outcome-based guarantees and proof of engagement, such measurement frameworks are becoming central to CTV planning and optimization. Attention-to-duration, in particular, is gaining traction as a way to understand not just whether an ad was viewable, but whether it was meaningfully watched.

For agencies navigating increasingly complex video buys across linear, streaming, and hybrid environments, the ability to validate attention at scale may become a competitive differentiator.

What This Signals for CTV Buying in 2026

Taken together, the findings suggest a shift in how CTV success will be defined going forward. Rather than defaulting to premium inventory assumptions, advertisers may increasingly prioritize data-powered audience strategies that deliver measurable engagement, especially as budgets tighten and accountability pressures rise.

As CTV continues to evolve into a results-driven channel, studies like this point to where the next layer of advantage may be found … not in where ads run, but in who they reach, and how closely viewers are actually paying attention.