Audience segments in India CTV are the targeting building blocks that define who a campaign reaches beyond simple geography. Unlike web advertising with its rich third-party cookie data ecosystem, CTV audience segments in India are derived from a narrower set of signals: device ID-linked browsing behaviour, publisher subscriber profiles, telecom-derived data, and DSP modelled audiences. Understanding how India CTV audience segments are built and their accuracy limitations is necessary for setting realistic targeting expectations.
How India CTV audience segments are built
DSP modelled audiences (DV360, TTD): The primary source for programmatic CTV audience targeting in India. DV360 builds audience segments from Google's identity graph — combining Android device IDs, Google account data, YouTube viewing signals, and search behaviour. The Trade Desk builds audiences from its Unified ID 2.0 and third-party data partnerships. These audiences are modelled — device IDs are assigned to segments based on behavioural signals, not declared data. Segment quality is strongest for broad, high-volume categories (18–34 age band, sports enthusiast, BFSI intent) and weaker for narrow or less common profile combinations.
Publisher first-party audiences: Platforms with mandatory registration — JioHotstar, Zee5, SonyLIV — collect self-declared subscriber data (age, gender, location, mobile number) that links to content consumption behaviour. Publisher first-party targeting is more accurate than DSP modelled audiences but operates within a single publisher's inventory. An advertiser targeting JioHotstar subscribers identified as "premium sports viewers" is using real first-party data for that publisher's impressions.
Telecom data partnerships: Jio subscriber data enriches JioHotstar's audience intelligence. Theoretically, Jio's telco data — app usage, location history, content consumption on Jio network — is one of the richest identity graphs in India. In practice, the level of data activation visible to advertisers is constrained by privacy considerations and is available only through JioHotstar's managed campaigns, not open programmatic.
Standard India CTV audience segment categories
| Category | Examples | Data quality in India |
|---|---|---|
| Demographic | M 25–34, F 35–44, SEC A household | Medium — modelled, not panel-verified |
| Geographic | Mumbai, Karnataka, Tier 2 cities | High — IP-based geo is reliable at city level |
| Content affinity | Cricket viewer, drama viewer, news viewer | High — derived from direct content consumption |
| Purchase intent | Auto intender, smartphone buyer, travel planner | Medium — stronger for high-volume categories |
| BFSI and financial | Insurance prospects, mutual fund considerers | Medium-low — India financial data is thinner |
| First-party (advertiser CRM) | Existing customers, lapsed users | Low match rate (15–35%) but high accuracy for matched users |
Audience segment limitations in India CTV
Device ID coverage gap: Approximately 30–40% of India CTV impressions lack a valid device ID. Audience segments only apply to the fraction of impressions with resolvable device IDs. A campaign targeting a 5 million device ID segment may reach only 3–3.5 million actual CTV device IDs after ID coverage is applied.
Cross-device identity fragmentation: India CTV audience segments built on Android TV device IDs do not automatically link to the same household's mobile identity, limiting cross-device frequency management and attribution.
Segment freshness: DSP modelled audience segments in India CTV have variable refresh cycles. Segments based on search and browsing behaviour may be 7–30 days stale, meaning "auto intenders" may already have purchased before the CTV campaign reaches them.
India audience segment best practice
Layer audience segments rather than relying on a single signal. A recommended stack for India CTV: geographic foundation (city or state) + broad demographic (age/gender band) + content affinity (relevant category) + behavioural intent (category-specific intent segment). This combination maximises the fraction of impressions that have valid targeting signals while filtering the most irrelevant inventory.