CTV Identity Crisis: 78% Fail in 2026

That 78% of advertisers figure is a real gut punch, it means they think their identity tools for CTV are failing them, and what’s worse, that number jumped 15% in just a year. The industry’s past approaches are clearly not cutting it for how brands need to reach people now, post-cookie. The only way forward for CTV identity is through genuinely new approaches built for this specific channel.

The Rise of First-Party Data: 60% of Brands Prioritize Direct Relationships

The numbers finally confirm what we’ve all been seeing on the ground: 60% of brands are pushing hard to collect and use their own first-party data, especially for CTV. This whole movement is being forced by the death of the third-party cookie and the intense focus on consumer privacy. I see it every day with the large retail brands I work with. The ones who invested years ago in a solid CRM and got their consent management right are in a great position. They aren’t scrambling. They’re building on what they already have.

So what’s the practical application for CTV identity? It means advertisers are finally ditching their reliance on shady data brokers and working to understand their own customers directly through website interactions, app usage, loyalty programs, and purchase history. The big stumbling block for most, though, is actually getting that data piped into the CTV world. Onboarding this data is genuinely hard. This is why data clean rooms, like the ones from AWS Clean Rooms or Snowflake, are becoming mission-critical. These platforms let advertisers match their customer lists against a publisher’s audience without either side seeing the other’s raw PII. Your first-party data is basically useless for CTV if you don’t have this kind of secure plumbing. Brands that cheap out on these data collaboration tools simply won’t be able to reach their best customers on the biggest screen in the house, and they will fall behind.

Contextual Targeting’s Comeback: A 40% Increase in CTV Adoption

While everyone was panicking about cookies, contextual targeting quietly made a massive comeback, with a 40% increase in CTV adoption in the last two years. Today’s contextual tech uses advanced NLP and machine learning to analyze the actual themes and sentiment of a show, looking at dialogue, visuals, and metadata. It’s not just basic keyword matching anymore. Think about it: an ad for a new EV appears during a documentary on climate science, or a DoorDash-type ad runs during a competitive cooking show. The relevance is baked right into the content, which means you don’t need to lean so heavily on tracking individual users.

People often write off contextual as being less precise than behavioral targeting, and sure, it won’t give you the same micro-level segmentation you once got with cookies. But the privacy angle is a huge win. There’s no PII, no cross-site tracking, just a smart match between the ad and the content. On CTV, where you’re often targeting a whole household anyway, this kind of relevance can be incredibly effective. Publishers are all over it because it’s a way to make money while respecting privacy. I’ve run campaigns on platforms like Magnite or PubMatic where a smart contextual strategy, layered with some basic geo-demographics, performed as well or even better than old behavioral campaigns. Transparent relevance, it turns out, can be just as strong as creepy hyper-personalization.

The Interoperability Imperative: Publishers Report 50% Higher CPMs for Unified IDs

The fractured CTV field, different platforms, different standards, different IDs, is a constant source of pain. But one stat tells you everything you need to know about the way forward: publishers who use unified, open-source IDs are seeing 50% higher CPMs than those stuck on proprietary tech. That economic pressure is what makes interoperability so important for CTV identity. Advertisers want to buy reach and measure it consistently, not get stuck inside a single walled garden.

The goal isn’t for one single ID to rule them all. The goal is to get different identifiers to actually talk to each other across the whole media buy. That’s why solutions like Unified ID 2.0 (UID2) are getting so much attention. They create a common language. UID2 works by taking an email address or phone number, encrypting it (hashing and salting it to protect privacy), and turning it into an ID that can be used for targeting, but only after the user has given consent. For publishers, the math is simple: better addressability brings in more ad dollars. For advertisers, it means you waste less money and get a clearer picture of campaign performance. The real work is getting everyone to adopt these IDs and integrate them into their tech stacks. I’m seeing ad tech vendors move fast to build in UID2 support because they know that a fragmented identity market is bad for everyone’s business. Money talks, and the industry is finally moving toward shared standards to chase those higher CPMs.

Measurement Gaps Persist: Only 35% of Advertisers Confident in CTV Attribution

Even with all the money pouring into CTV, a massive problem remains: only 35% of advertisers are confident they can actually measure ROI from their campaigns. That number is frankly terrifying. It shows just how broken our measurement models are for a world without cookies. The last-click attribution models we built for desktop just fall apart when you’re dealing with a family watching a show on a TV and then someone buying the product on their phone hours later. How do you connect those two events without a persistent ID?

This uncertainty isn’t a small problem. It messes with budgets and cripples strategic planning. Advertisers have to get past these simplistic models and start using more grown-up approaches like multi-touch attribution, properly structured incrementality tests, and geo-lift studies built for the realities of CTV. Measurement companies like Nielsen One or Innovid are shipping solutions that try to connect the dots using panel data and probabilistic models, but the whole field is still very new. I tell my clients to use a mix of tactics instead of waiting for a single magic bullet. Test different attribution windows, run A/B tests in different markets, and connect your CTV exposure data to real business results (not just clicks). The insights won’t be handed to you, you have to invest in the analytics to go find them. If you don’t, even the best identity solution on the market will look like a failure.

The Enduring Power of Device Graphs: 25% of Ad Spend Still Relies on Probabilistic Matching

For all the talk about first-party data and deterministic IDs, a quarter of all CTV ad spend, 25% of it, is still being targeted using probabilistic device graphs. This fact shows the real-world complexity of identity and why you need a balanced strategy. Probabilistic matching is essentially making an educated guess, linking devices based on shared signals like an IP address, Wi-Fi network, or location data. It’s not a 100% certain link, but with good data and clean algorithms, it’s an effective way to get reach, particularly when you need to find new customers or fill in the gaps where you don’t have deterministic matches.

The common knock against probabilistic methods is that they’re less accurate and worse for privacy. I don’t entirely buy that. Deterministic matching is great for confidence but often gives you very limited scale. A well-managed probabilistic graph from a reputable company like LiveRamp or Zeotap can massively extend your reach. The quality of the underlying data sources and the sophistication of the matching algorithms are what matter. And the privacy side can be managed responsibly when the graphs are built with anonymized data and give users clear ways to opt out. For a lot of brands that don’t have a mountain of first-party data, these graphs are an absolutely necessary piece of their CTV identity strategy. To just ignore them is a major strategic mistake.

The move to cookie-less CTV isn’t one problem with one solution. It’s a complex challenge that demands a smart combination of first-party data activation, modern contextual targeting, interoperable IDs, and far more sophisticated measurement. The brands that get their heads around this and build a flexible strategy are the ones that are going to win.