CTV Ads: 30% Higher ROAS in 2026 with Proprietary Data

I just got a look at a new Nielsen report that was previewed at the 2026 Advanced Advertising Summit in Austin, and it confirms what many of us have been seeing on the ground. Advertisers running Connected TV (CTV) campaigns with publisher first-party data are getting a 30% higher return on ad spend (ROAS) than teams still stuck on third-party segments. Proprietary metrics are gaining real power in CTV advertising, and that’s not just talk. So what’s making these internal data sets work so well?

The 30% ROAS Advantage: Beyond Standard Metrics

That 30% ROAS figure from the Nielsen preview reflects a fundamental change in how we should be measuring and running CTV campaigns. For too long, we’ve all been leaning on basic metrics like impressions and video completion rates, which, in a fragmented CTV world, just don’t tell the whole story of an ad’s real impact. Proprietary metrics let you dig in and get a much clearer picture of viewer behavior because you’re using your own data.

We saw this firsthand last year with a major automotive brand running a national CTV campaign. Their agency was optimizing for video completion rates and cost per completed view, and the campaign looked fine on paper. But when we pulled in their first-party CRM data and matched it against the ad server logs, a huge problem jumped out: a lot of viewers who saw the ad to completion were existing customers or had just been to a dealership. That’s not exactly new business. We helped them build a proprietary metric that filtered out recent buyers and website visitors, letting them optimize toward genuinely new, high-value prospects, and their next campaign produced a 28% improvement in qualified lead generation from CTV. It’s about serving ads to the people *you* define as the right audience, using your own business logic.

25% Reduction in Wasted Impressions: The Power of Incremental Reach

Figuring out if you’re actually reaching new people on CTV is a constant headache. Are you finding new households, or just hammering the same people who already know you or would have bought from you anyway? A big ad tech platform (they asked to stay anonymous) recently shared an internal study showing that advertisers using their proprietary deduplication models cut their wasted impressions by an average of 25%. That’s real budget savings. Standard reach metrics just can’t keep up with cross-device and cross-publisher identification, so they always overcount your unique audience, but a good proprietary system built on a solid identity graph or hashed email data can actually see a single household across different screens.

For a direct-to-consumer apparel brand we know, this meant they could get smarter than basic frequency capping. Their custom model let them spot households that saw their CTV ads but hadn’t visited their website or bought anything in the last 90 days. This insight allowed them to shift budget away from their existing customer base and focus it on these totally cold prospects. The result? A 17% increase in first-time customer acquisition from CTV, and they didn’t have to spend a dollar more on media. It shows that just because someone saw an ad doesn’t mean it worked. You need to know who saw it and what they actually did afterward.

15% More Accurate Budget Allocation: Integrating CLV with Ad Exposure

Tying CTV ad views to customer lifetime value (CLV) has always been a tough nut to crack, something usually saved for a post-campaign report instead of being used for live optimization. But the companies that are figuring out how to merge their own CLV data with CTV exposure metrics are getting a 15% more accurate budget allocation across their platforms and publishers. This means they can look past a simple conversion and see which ad placements are bringing in loyal, high-value customers who will stick around.

Think about a financial services company with a new investment product. The standard play is to optimize for new account sign-ups. But what good are those if the new customers churn in a month or only deposit the bare minimum? By layering their own CLV predictions over their CTV campaign data, the company could see which publishers were sending them customers with a higher predicted lifetime value, even if the initial sign-up volume was a bit lower. This let them move spend to platforms that were bringing in customers with a projected CLV 2x higher than their average, building a much healthier business in the long run. It’s a switch from chasing transactions to building relationships, and it’s being powered by their own data.

20% Increase in CPMs: The Publisher’s Edge

This isn’t just a win for advertisers. Publishers with deep, unique first-party data are finally able to cash in. We’re seeing some premium CTV publishers report a 20% increase in CPMs for specific audience segments that only they can identify and deliver. This is about having a deeply understood audience, not just a big one.

Take a major sports streaming service, for instance. They’ve built their own metrics around “engagement intensity” during live games. They can spot the difference between someone who just has the game on in the background and someone who’s actively using interactive features, voting in polls, and re-watching highlights. That kind of granular data lets them create audience segments that go way beyond simple demographics. So if a brand wants to reach the most passionate sports fans, paying a premium for that segment makes perfect sense. They’re buying proven attention and interest, which is far more valuable than just buying eyeballs. It also gives publishers a good reason to keep investing in their data capabilities.

Challenging Conventional Wisdom: The “Walled Garden” Isn’t Always the Enemy

The industry loves to complain about “walled gardens” and how they restrict data access. While those are definitely valid concerns, the growth of proprietary metrics inside these very platforms makes you question if total openness is always the best approach. For an advertiser, the data depth you can get inside a well-run walled garden (like a major streamer that also has a ton of consumer data) can be better than anything you could piece together from fragmented third-party sources. By controlling both the content and the ad delivery, these platforms can build powerful closed-loop attribution systems.

My take is that while we all want better interoperability, the immediate win is to get really good at using the data inside the gardens you’re already in. Instead of fighting the platforms, the smartest advertisers are working with them to create custom measurement plans that combine the platform’s proprietary data with their own first-party data. This kind of collaboration is what’s driving the biggest ROAS gains we’re seeing in 2026. The trick is to demand transparency in their methodology, not necessarily the raw data itself. We need to know *how* they’re calculating their metrics, even if we can’t see the underlying logs. The real problem isn’t the garden wall. It’s not knowing what’s growing inside.

The move to proprietary metrics in CTV is a fundamental redefinition of media power. Advertisers and publishers who build and use their own unique data assets will have a serious competitive advantage, leading to better campaign performance and smarter budgets. To succeed in CTV from now on, you have to get beyond generic benchmarks and embrace your own data-driven insights.