Audience targeting in CTV advertising has matured from a largely contextual exercise — targeting genre and daypart because nothing better was available — into a multi-signal system capable of reaching specific household profiles with near-deterministic precision. The shift happened because the television ecosystem developed identity infrastructure that did not exist in linear TV: platform advertising IDs on streaming devices, subscriber data from streaming services, ACR viewing history from smart TV operating systems, and household graphs that resolve all of these signals to a persistent household key. The result is that the audience targeting available in CTV programmatic is in many respects more precise than what was possible in digital display advertising at its peak cookie-based sophistication — operating at the household level rather than the individual device level, with television viewership context that no web signal can provide. This guide maps the complete CTV audience targeting landscape: what segments exist, how they are built, their accuracy profiles, and how to configure them effectively through the LtvAdx advertiser platform.
The audience data stack in CTV
CTV audience data comes from four source categories arranged in a quality hierarchy. First-party advertiser data — CRM files, email lists, loyalty program members — sits at the top of the hierarchy because it is owned by the advertiser, deterministic, and directly tied to known customer behavior. Second-party publisher data — registration information from streaming service subscribers — is the next tier, offering near-deterministic accuracy for publishers with large authenticated user bases. Third-party data — modeled demographic, behavioral, and purchase intent segments from data brokers — provides broad reach at lower precision. ACR data from smart TV manufacturers occupies a unique position: it is third-party in origin but uniquely CTV-native because it captures actual television viewing behavior that no other data source provides.
The LtvAdx HouseholdID system operates as the resolution layer that connects these data sources to actual impressions. When a bid request arrives from a Roku device, the HouseholdID graph resolves the device's RIDA (Roku advertising ID) to a household key, checks what audience segments have been matched to that household key, and passes the segment membership to the auction logic so campaigns with matching targeting criteria can bid. This resolution happens in under 10ms at ad decision time, within the total auction timeout budget for the break. The CTV identity without cookies guide covers the technical mechanisms of this resolution in depth.
First-party CRM audience segments
First-party CRM audience activation is the highest-value targeting approach available in CTV programmatic. The mechanics: the advertiser exports a customer or prospect list as hashed email addresses (SHA-256 or MD5), uploads the file to the DSP or data collaboration platform, and the system matches the hashed emails against the HouseholdID graph to produce a set of matched household keys. Campaigns targeting the matched household segment will serve ads to streaming devices that resolve to those households.
Match rates vary by advertiser type and list quality. Consumer brands with large direct-to-consumer email programs typically see 45–65% match rates against major household graphs because their customer emails are widely used for streaming service sign-ups and retail loyalty programs. B2B advertisers with work email-dominant contact lists see lower match rates — 20–35% — because work emails rarely appear in consumer streaming service registrations. Premium match rates (65–80%) are achievable when the advertiser's customer list includes mobile phone numbers alongside emails, because phone number matching adds a second resolution path.
First-party segments are valuable for three distinct use cases: customer retention (serving ads to existing customers to drive repeat purchase or upsell), suppression (excluding existing customers from acquisition campaigns to avoid wasted spend), and lookalike expansion (using the matched household set to build a model of similar households for prospecting). Configure all three uses of your CRM data in the LtvAdx advertiser platform before launch — suppression in particular is frequently omitted and results in measurable waste on acquisition campaigns that reach existing customers.
Demographic segments: income, age, and household composition
Demographic targeting in CTV draws from two primary sources: operator subscriber data from MVPDs and vMVPDs who have household-level demographic records, and modeled demographic segments built by data providers from multiple signal sources including credit bureau data, retail loyalty programs, and voter registration records. Operator-sourced demographics are more accurate because they are based on billing records; modeled demographics are more broadly available across the full CTV supply universe.
Income targeting is particularly valuable in CTV because the platform's household-level identity maps cleanly to financial data. Targeting households in the top income quintile (household income above $100,000) for luxury automotive, financial services, and premium travel is a high-value CTV use case that outperforms comparable digital targeting because IP-to-household matching provides stronger income signal than individual-level cookie-based inference. Household composition segments — families with children, empty nesters, multi-generational households — are similarly accurate in CTV because billing and subscriber data captures household structure rather than individual-level inference.
Age-based targeting in CTV should be calibrated for household-level behavior: a household classified as "adults 25–54" may include multiple adults of different ages who share a single streaming account. CTV age targeting is best understood as household age profile targeting rather than individual-level age targeting. For campaigns requiring precise individual age targeting — alcohol brand age verification, pharmaceutical advertising — CTV household-level targeting requires additional age-gate controls at the campaign level rather than relying on segment accuracy alone.
Purchase intent and in-market segments
In-market audience segments identify households that have exhibited behavioral signals indicating active purchase intent in a specific category: recent vehicle configurator visits for auto, mortgage pre-qualification searches for financial services, travel search and booking behavior for hospitality. In digital advertising, these signals come from web browsing data via cookie or device ID; in CTV, they come from third-party data providers who match retail and search behavior to household identifiers via IP.
The accuracy of in-market CTV segments is a function of how recently the behavioral signal was observed and how deterministic the household identity match is. A household classified as "in-market for a new SUV" based on web searches from three months ago may have already purchased — segment freshness is a significant quality variable. Evaluate third-party in-market segment providers on two dimensions: recency of signal update (weekly updates are minimum; daily is better) and identity match methodology (deterministic IP match preferred over probabilistic device graph extension).
For performance-oriented campaigns, in-market segments are most effective when layered with content genre targeting rather than used alone. A household in-market for a mortgage who is watching financial news content on a FAST channel is a higher-intent target than the same household watching crime drama — the content affinity amplifies the behavioral signal. Configure combined targeting in the advertiser platform using AND logic between in-market segment and genre targeting to capture this signal combination. The CTV for performance advertisers guide covers the full performance targeting strategy including lookalike expansion and incrementality measurement for in-market campaigns.
ACR-based audience segments
Automatic Content Recognition data from smart TV manufacturers — Samsung, LG, Vizio — creates a uniquely valuable targeting layer available only in CTV. ACR embeds in the TV operating system and identifies what the television is displaying by fingerprinting audio and video frames against a reference database. This creates verified viewing history: not modeled or inferred content affinity, but a record of what that specific television screen actually displayed, including linear cable content, broadcast antenna programming, gaming console output, and OTT streaming apps.
ACR audience segments serve three distinct advertising purposes. Competitive conquesting targets households that have recently been exposed to a competitor's advertising on linear TV — the ACR signal identifies the competitor's commercial by fingerprint recognition. These households are reached with the advertiser's message on CTV streaming apps running on the same smart TV, creating a direct competitive response in the same household environment. Linear TV reach extension identifies heavy linear TV viewers — households that watch 30+ hours of linear TV per week — who may have limited streaming app usage and are reached through linear addressable inventory from MVPD operator partners. Genre-based behavioral segments use ACR viewing history to classify households by content preference: verified sports fans, home improvement content enthusiasts, documentary viewers — based on actual viewing rather than self-reported or modeled interests.
Contextual targeting: genre, content, and channel
Contextual targeting in CTV operates without identity signals: the ad is selected based on the content being watched rather than who is watching it. Genre targeting (sports, news, entertainment, kids) is the most commonly used contextual dimension, implemented by mapping publisher-declared content categories in the OpenRTB bid request to campaign eligibility criteria. Content-level targeting — specific series titles, specific channels — is available through private marketplace deals with publishers who expose title-level metadata.
Contextual targeting is the appropriate fallback when identity match is unavailable (device in LAT mode, opted-out household, kids content restrictions) and the primary targeting strategy for campaigns where content adjacency is the core value proposition — a cooking brand advertising in food programming, a sporting goods brand in live sports. For the brand safety dimension of contextual targeting, the CTV brand safety guide covers how IAB content categories enforce adjacency restrictions.
Geographic and DMA-based targeting
Geographic targeting in CTV ranges from national reach down to zip code precision. DMA-level targeting (Designated Market Area — the 210 geographic TV advertising markets defined by Nielsen) is the standard for television advertising geographic planning and is fully supported in CTV programmatic. IP-to-DMA mapping in CTV is highly accurate for residential broadband households — typically 95%+ DMA match accuracy for fixed broadband connections.
Below DMA, zip code targeting enables retail advertisers, local service businesses, and regional campaigns to reach specific community-level geographies. Zip code precision relies on the same IP-to-household matching that powers demographic targeting, with similar accuracy profiles: strong for residential fixed broadband, weaker for mobile internet and shared IP environments. For campaigns requiring precise local targeting — a regional car dealer targeting zip codes within 30 miles of their dealership — CTV zip targeting combined with a household income overlay is a strong alternative to linear local cable buying, offering better reporting and identity-linked frequency control.
For political advertising specifically, district-level geographic targeting using congressional district, state senate district, or county-level boundaries is available through specialized data overlays. Political CTV targeting is covered in the broader context of the regulatory and compliance framework in the CTV political advertising guide.
Building audience strategy: layering and scale tradeoffs
The temptation in CTV audience configuration is to apply every available targeting dimension simultaneously — first-party CRM, in-market segment, income tier, genre, and geographic — and achieve the most precise possible audience definition. The problem is that each additional targeting layer reduces available inventory volume exponentially. A campaign targeting in-market auto shoppers (say, 8% of CTV households), in the top income quartile (25% of households), in specific DMAs (say, 40% of national inventory), watching automotive or news content (15% of total viewing time), produces an addressable audience that may be too small to pace effectively.
A practical framework: start with two or three targeting dimensions that define your core audience clearly, and run the campaign for the first week to establish baseline delivery and performance. Add a fourth dimension in week two only if performance data suggests it will improve outcome rates without destroying scale. Use the LtvAdx reporting audience reach estimates before campaign launch to validate that your targeting combination produces a household universe of sufficient size to deliver your impression goal within the flight window. For campaigns with aggressive reach targets, build an audience expansion fallback tier into the line item structure — broadening to contextual-only targeting for impressions that cannot find a segment match — rather than leaving pacing gaps when premium segments are exhausted. To discuss audience strategy for a specific campaign, request a platform walkthrough with the LtvAdx advertiser team.
