How Identity Resolution Works for CTV Audience Targeting

Identity resolution for connected TV works by linking fragmented device signals (streaming device IDs, IP addresses, and hashed emails) into a unified household profile. That profile is what CTV platforms use to match your ads to specific people rather than demographic approximations. It’s the mechanism that separates a performance channel from a reach channel.

On Meta and Google, this layer is invisible — those platforms built their identity systems around logged-in users over years. On streaming TV, the same infrastructure exists, but the mechanics are less familiar. Understanding how it works makes CTV much easier to use like a performance marketer.

What is identity resolution in CTV advertising?

Identity resolution in CTV is the process of connecting a person’s fragmented digital signals into a single, actionable profile. A streaming TV device (a smart TV, Roku, or Apple TV) doesn’t carry cookies. There’s no persistent identifier that travels between apps or platforms the way a browser cookie does. Identity resolution solves this by using other signals to identify which household or person is watching, then links them to their behavior elsewhere.

Two approaches do the matching. Deterministic matching uses verified, known data (a hashed email from a streaming app login, a subscriber record, a first-party CRM upload) to create a confident one-to-one connection between a device and a real person. Probabilistic matching takes a wider approach, clustering devices that share the same IP address, show similar usage patterns, or consistently co-locate, and inferring they belong to the same household.

Deterministic is high confidence but requires a data source with a verifiable identifier. Probabilistic extends coverage where deterministic signals don’t exist, but introduces uncertainty. The most accurate CTV identity graphs use deterministic signals at their core and probabilistic signals to fill the gaps.

Without identity resolution, a CTV impression lands in a household but can’t be connected to a specific person, their purchase history, or what they do after seeing the ad. With it, that same impression becomes the foundation for retargeting, attribution, and audience segmentation.

How does a CTV identity graph actually work?

A CTV identity graph is a database that links together all the signals that point to the same household or person (streaming device IDs, IP addresses, hashed email addresses, and mobile device identifiers) into a unified record. When a CTV platform targets a campaign, it queries that graph to find which households match your audience criteria and which can be reached via streaming inventory.

The graph operates at two levels: household and individual. Most CTV targeting runs at the household level, because a TV is a shared device. A household-level match links the TV to the home’s IP address, then to the other devices (phones, laptops) operating under the same roof. This is what makes cross-device attribution work: a TV ad impression connects to a web visit that happened later because both events resolve to the same household record.

Scale determines precision. A larger identity graph means broader household coverage, more accurate segment matching, and better lookalike modeling. For CTV audience targeting to function like a performance channel, the identity layer needs to be deep enough to match your first-party data against real, reachable households with meaningful behavioral signals attached.

Each resolved profile enables specific targeting actions: retargeting a site visitor, suppressing existing customers from acquisition campaigns, extending a seed audience into lookalikes, or controlling how often a specific household sees an ad across a campaign.

How do you connect your first-party data to CTV targeting?

CTV without first-party data matching isn’t performance TV. It’s brand advertising with a shorter production timeline.

Demographic targeting (age, income, content genre) is reach targeting. It doesn’t produce the signal precision that performance marketers get from Meta and Google, because it doesn’t use what you already know about your customers. First-party data matching closes that gap.

On Vibe.co, activating your first-party data works the same way as audience activation on your existing paid channels. You upload a CRM list or sync directly from tools already in your stack: Klaviyo, HubSpot, Shopify, and others available through the integrations marketplace. Vibe’s Identity Intelligence then matches that list against a 120M+ household profile graph to find the streaming households you can reach. The same audience segments you built in Klaviyo (non-purchasers, lapsed customers sorted by LTV, post-purchase upsell cohorts) activate directly on streaming TV without rebuilding anything. See the full audience targeting features for detail on what the match layer supports.

Knix, the DTC apparel brand, connected their full Klaviyo segmentation directly to Vibe. Non-purchasers, lapsed customers organized by lifetime value, and lookalike audiences built from high-value buyers all ran as live targeting inputs on streaming TV. The result was 5.6x ROAS during their April sale period, verified by Northbeam, with 3–4x consistent ROAS month over month. “You guys have gone out saying you’re the social ads manager of TV,” their team noted. “That’s exactly what it is.” The Knix case study details how they structured the segmentation across funnel stages.

Vibe earned the G2 Best Estimated ROI award in the Mid-Market category — a signal that what comes out of the identity matching layer is producing verifiable outcomes, not just impressions.

What audience types can CTV identity resolution power?

First-party data matching powers four distinct audience types on streaming TV, each using the resolution layer differently.

Branded Bills, the custom headwear brand, ran IP matching through their Klaviyo integration, connecting CRM data to streaming households via the identity graph. The campaign delivered 311% ROAS overall, with a peak retargeting day reaching 2,209% ROAS. A similar pattern appears in the Sijo case study: Klaviyo segmentation activated through Vibe delivered 304% ROAS and 57% lower new customer CAC versus social, verified by Northbeam.

How does identity resolution affect CTV measurement and attribution?

Measurement on connected TV runs through the identity layer. Without resolved household profiles, you can’t connect a TV impression to an action that happened on a different device.

When someone sees your streaming ad and then visits your site on their phone 20 minutes later, attribution works through identity resolution: the TV impression gets tied to a resolved household profile, the phone visit gets tied to the same profile, and the conversion gets attributed to the campaign. Without that resolution step, those two events look unrelated.

Holdout-based incrementality testing uses the same identity layer. The test group (households exposed to the ad) and the holdout group (matched households that weren’t) are defined at the resolved household level — which means the comparison is clean. Identity resolution accuracy and measurement accuracy are the same problem. Vibe’s CTV measurement integrations with Northbeam, Triple Whale, and Haus work because event-level data carries the identity signals needed to connect TV impressions to downstream conversions across devices.

One practical caveat worth noting: match rates depend on your CRM data quality. A clean, deduplicated email list matches against the identity graph at significantly higher rates than a stale one. Probabilistic matching extends coverage where deterministic signals aren’t available, but it introduces inherent uncertainty. The cleaner your first-party data going in, the more precise your targeting and the more accurate your measurement coming out.

Frequently asked questions

How does identity resolution work for CTV audience targeting?

Identity resolution for CTV works by linking fragmented device signals (streaming device IDs, IP addresses, hashed email addresses) into a unified household profile. CTV platforms use that profile to match ads to specific audiences, enable cross-device retargeting, and attribute conversions back to TV impressions. Without it, CTV targeting is limited to demographic approximations rather than specific, identifiable people.

What is a CTV identity graph?

A CTV identity graph is a database that links all the signals identifying a household or individual (streaming device IDs, IP addresses, email hashes, mobile device identifiers) into a single unified record. Advertisers query the graph to find reachable households that match their audience criteria. The scale and quality of the graph determine how precisely you can segment and how accurately you can attribute results.

How do I target my existing customers on connected TV?

Upload a CRM list or sync from an integration like Klaviyo, HubSpot, or Shopify, and the CTV platform matches it against the identity graph to find reachable streaming households. On Vibe, this works the same way as first-party audience activation on your existing paid channels — the same segments you built in email or paid social activate directly on streaming TV without rebuilding anything.

Does CTV audience targeting work without third-party cookies?

Yes. CTV identity resolution uses deterministic signals (hashed emails, subscriber logins, first-party CRM data) rather than third-party cookies. Streaming TV was never cookie-based, so cookie deprecation doesn’t affect CTV targeting the way it affects open-web display. First-party data matching and household IP resolution both operate independently of cookies.