Panel measurement and census measurement are the two fundamental approaches to counting TV audiences. In linear TV, panels are the only option — you sample a set of households and extrapolate to the total population. In CTV, census measurement is possible because every impression is tracked at the server level. Understanding which approach your measurement vendor or platform uses — and what each approach gets right and wrong — is essential for any serious CTV buyer in India.
What is panel measurement?
Panel measurement works by recruiting a representative sample of households, installing measurement devices in their homes, and recording what those households watch. The data from this sample is then statistically extrapolated to represent the total viewing population.
BARC India (Broadcast Audience Research Council) operates India's primary TV measurement panel. As of 2024, BARC's panel covers approximately 50,000 homes across urban and rural India, extrapolated to represent a universe of roughly 900 million TV viewers. Every BARC rating point — the GRP that linear TV buys are transacted on — comes from this panel.
How BARC's panel works
BARC installs a device called a BAR-O-Meter in panel homes. This device uses a combination of audio watermarking and set-top box integration to detect which channel is being watched and by whom (panel members press buttons on a remote to register their presence). The data is aggregated, weighted, and published as ratings for channels and programmes.
Panel measurement has been the backbone of Indian TV buying for decades. It is the currency for linear TV transactions — every GRP, reach, and frequency figure in a linear TV plan is panel-derived.
Strengths of panel measurement
- Cross-platform and cross-channel: A panel measures all TV content reaching a household — DTH, cable, IPTV, and CTV — from a single source. This is the only way to get genuinely unduplicated reach across TV environments.
- Person-level data: Panels can capture which individual in the household was watching (if members register themselves). This allows demographic breakdowns that platform-level data often cannot provide.
- Consistent currency: Because all channels are measured by the same panel methodology, the ratings are directly comparable. You can legitimately compare Star Plus GRPs to Sony GRPs.
Weaknesses of panel measurement
- Sample size limitations: A 50,000-home panel extrapolated to 900 million viewers means each panel home represents roughly 18,000 actual households. Small sub-groups — regional language viewers, specific city-level audiences, niche content genres — can have statistically unreliable ratings.
- Slow reporting: Panel data is typically available with a day or more of lag. Real-time optimisation during a campaign is not possible.
- Co-viewing assumptions: Panels estimate who is in the room through button-press registration. Non-registrations are common. Co-viewing estimates are imprecise.
- Recruitment and attrition: Maintaining a representative panel is operationally hard. Panel homes change behaviour over time (the "panel effect"), and low-income or rural homes are harder to recruit and maintain.
What is census measurement?
Census measurement means tracking every impression served — not a sample, but the complete population of ad delivery events. CTV enables this because every ad request, impression, and completion event is logged by the ad server or streaming platform at the individual device level.
When a pre-roll ad plays on JioCinema via a connected TV, the platform's ad server records: which device, which geographic location (IP-based), what time, what content was playing, and whether the ad was completed. This is census data — 100% of impressions, not a projection from a sample.
Strengths of census measurement
- Complete impression count: No extrapolation required. You know exactly how many times your ad was served, not an estimate.
- Real-time or near-real-time: Census data from ad servers is available within hours, sometimes minutes. This enables mid-campaign optimisation.
- Geographic and device granularity: You can see impression delivery broken down by city, state, device type, and operating system — data that panel measurement cannot provide at this granularity.
- Frequency management: Because every impression is tracked, frequency caps can be enforced at the user or device level. This is impossible with panel data.
Weaknesses of census measurement
- Platform siloing: Census data from JioCinema tells you about JioCinema viewers. Census data from SonyLIV tells you about SonyLIV viewers. These datasets do not automatically connect. A viewer watching both platforms will appear as separate devices in each platform's census unless a shared identity graph links them.
- No person-level data: Census data measures device-level impressions, not people. A CTV device shared by four family members generates four people's viewing but appears as one device ID. Co-viewing is not captured.
- Publisher-reported data: When census data comes directly from the platform selling you the inventory, there is an obvious conflict of interest. Third-party census measurement (through a verification vendor like DoubleVerify or IAS) addresses this but is not universal in India.
- No content-neutral currency: Unlike BARC ratings which cover all TV, census measurement is ad-delivery measurement. You measure ad impressions, not total audience for a content piece.
How panel and census data are combined in practice
The leading approach in markets with more mature CTV measurement — and the direction India is heading — is to combine panel and census data through a process called calibration or fusion.
The idea: use the census data's complete impression count for scale and granularity, then calibrate it against panel data for demographic composition and person-level estimates. Nielsen ONE in the US uses this approach — census-level digital data calibrated against a panel to produce person-level reach and frequency estimates. BARC India and some India platform data providers are exploring similar methodologies.
The practical challenge: effective calibration requires data linkage between the panel (which measures a sample of devices and people) and the census data (which measures devices but not people). In India, this linkage is still at an early stage.
India context: BARC panel vs OTT platform data
India currently has a significant measurement gap. BARC India's panel is the authoritative source for linear TV — every TV plan and post-campaign report uses BARC data. But BARC's CTV and OTT measurement capability is limited. BARC's streaming measurement (called EKAM, now in development) aims to bring OTT platforms into a unified measurement framework, but as of early 2026 it is not yet universally adopted.
What this means in practice:
- If you run a campaign on JioCinema CTV, your delivery data comes from JioCinema's ad server (census) — but there is no BARC panel validation of that data.
- Cross-platform reach — "how many unique people did my campaign reach across JioCinema CTV and SonyLIV CTV?" — cannot be answered using either panel or census data alone. It requires a shared identity infrastructure that does not yet exist at scale in India.
- Agencies running blended linear + CTV campaigns must aggregate linear BARC data with platform-reported CTV impression data — two methodologically different datasets that cannot be directly compared.
For India CTV buyers today, the practical advice is: treat platform-reported census data as your campaign delivery truth for CTV, and use BARC data for linear TV. Do not conflate the two into a single "reach" number without a clear methodology for deduplication.
What advertisers should ask
When evaluating measurement for a CTV campaign in India, ask these questions:
- Is this panel-derived or census data? Know which you are looking at before drawing conclusions.
- If census: is it third-party verified or self-reported by the platform? Third-party verification (DoubleVerify, IAS) adds credibility to impression counts.
- What is the identity methodology? How are repeat viewers identified across sessions? Is it deterministic (login-based) or probabilistic (modelled)?
- How is co-viewing handled? If the platform claims person-level reach on CTV, ask specifically how they account for multiple viewers per device.
- Is cross-platform deduplication possible? If you are running across multiple CTV platforms, understand that unduplicated reach is not directly measurable without an agreed identity graph.