Three years ago, most CTV campaigns ran on the same handful of levers—fixed audience segments, lookalike modeling, and manual bid adjustments based on last week’s performance report. If this describes your campaigns today, you are not alone. That is still the default across most of the industry.

What has changed is how much is riding on those rusty levers. U.S. CTV ad spend is on pace to hit $37.95 billion this year, up 14.5% year-over-year, per eMarketer. That means more budget is moving through CTV than ever, and most of it is still running through that same rules-based approach, now wearing an AI trench coat.

Cognitiv's deep learning CTV solutions work smarter, learning directly from campaign data in real time and getting more accurate with every impression served. Three pillars make that possible, each one earning its own section below.

Pillar 1: Precision

Boiled down, precise targeting in CTV means reaching households that are likely to resonate with your campaign. With interests and in-market signals shifting constantly, capitalizing on that resonance is more complex than it sounds. 

Think about your own household. Is what you searched for or considered buying a month ago still what you are in-market for today?

That constant shift is exactly why most CTV targeting falls short. It relies on fixed audience segments, broad categories that households get sorted into at the start of a campaign based on assumptions made before a single impression runs. Meaning, households continue to be targeted throughout the flight no matter how their behavior changes along the way.

Cognitiv's dynamic audience creation tool, AudienceGPT™, was built to solve this static audience problem. Trained on behavioral signals from over 250 million people, it moves households in and out of a segment as new signals come in. That means targeting that gets smarter the longer a campaign runs, rather than staying locked to day-one assumptions. 

A consumer electronics brand put AudienceGPT to the test in a recent CTV campaign during the World Cup.

Sports fandom shifts constantly during a tournament like this, driven by which teams are still in it, who's hosting the next watch party, and whatever storyline is dominating that week's matches. A fixed segment built before kickoff would miss most of that movement.

AudienceGPT identified the fans, hosts, and upgraders most likely to act as those signals emerged, continuously updating who counted as high-value rather than locking in a single definition for the whole tournament.

This ability to keep finding who mattered most led to a 64% lift in purchase intent and an 84% lift in ad awareness, measured through DISQO brand lift studies.

Pillar 2: Performance

Performance means something different for every advertiser. A CTV campaign can hit its reach goal, land every impression on premium inventory, post a strong completion rate, and still fail the one specific metric that actually matters to leadership. 

Ultimately, that is not three wins and one loss; it is just one loss.

Most CTV campaigns optimize toward reach and completion rate anyway, because those numbers are easy to chase the same way across every advertiser. 

Cognitiv's Custom Algorithms take a more intentional path instead, training a model on each advertiser's own first-party data, built around their specific KPI. That same model extends natively into CTV, rather than treating it as a separate, bolted-on channel.

A national cheese brand recently went the Custom Algorithm route on CTV to lift sales for a new refrigerated snack line. They aimed to prove that media could drive measurable sales and grow the base of households buying the brand. 

Cognitiv built a custom model on category purchaser data and real-time behavioral signals, then activated it across CTV and online video together, with creative tied to specific SKUs at the shelf.

The strongest lift, 62%, came from new and competitive buyers, meaning shoppers who had previously reached for rival brands at the store started reaching for this brand instead. 

Household sales also climbed 35.5%, 13x the category benchmark, with an incremental ROAS of $5.89.

Pillar 3: Brand Safety

Say a campaign nails its targeting, reaching exactly the households that are in-market, or its model is dialed into the exact KPI leadership cares about. Then the ad runs right after a segment about a violent crime...

Performance suffers regardless of how sharp that targeting or that model was, because the environment the ad landed in undid it.

That is the core of brand safety, and why it bears the most load of the three pillars. 

Precision finds the right household. Performance proves the right outcome. Neither survives contact with the wrong inventory, so without brand safety, precision and performance cannot support a smarter CTV campaign on their own.

Cognitiv approaches this by controlling where an ad can land before the campaign ever runs, packaging inventory with publisher and app blocklists and sourcing exclusively from top-tier CTV SSPs like Magnite.

That premium inventory footprint spans over 250 million households across publishers including Warner Bros. Discovery, Paramount, CBS, Fox, Roku, Samsung TV Plus, Fubo, Sling, and DIRECTV.

Brand safety ultimately protects whichever path a campaign runs on, precision or performance, by design, not leaving an ad's landing spot to chance.

Three Pillars, Every Campaign Covered

In practice, a single campaign builds around brand safety plus one other pillar, precision or performance, depending on what it is solving for. 

AudienceGPT sharpens who a campaign reaches as behavior shifts mid-flight, the pillar for a campaign built on reach and awareness. 

Custom Algorithms train a model around one advertiser's own KPI and first-party data, the pillar for a campaign built to move lower-funnel outcomes.

CTV budgets are only getting bigger. Advertisers who pair the right path, precision or performance, with real brand safety turn that growth into an advantage. Everyone else just absorbs more risk.

What could smarter CTV look like for you?

Get in touch with our team of CTV experts to learn how you can drive smarter CTV campaigns with Cognitiv.