Panel measurement recruits a statistically representative sample of households, installs measurement equipment in their homes, and uses that sample's viewing behaviour to project estimates for the broader population. The accuracy depends on the panel being demographically representative of the full population. Traditional TV audience measurement (BARC in India, Nielsen in the US) is panel-based — a panel of roughly 20,000–50,000 households is used to estimate what 100 million+ TV households are watching. Census measurement captures data from every single device or viewer without sampling — all impressions, all users. Digital advertising is largely census-measured: your DSP counts every impression actually delivered to every device.
For India CTV, the distinction matters because the two approaches have different strengths. Census measurement (from streaming platforms' own reporting and third-party ad verification) is highly accurate for delivery: how many impressions were served, VCR, which content was running. But census data alone does not tell you who was watching — device IDs are not people. Panel measurement provides the demographic overlay (age, gender, SEC) but projects from a small sample that may not adequately represent India's diverse CTV-owning population, especially for regional markets. The ideal is panel-plus-census: use census data at scale, calibrated by panel data for demographics. This is what Nielsen ONE attempts in the US. In India, no vendor offers this hybrid at scale for CTV in 2026 — the ecosystem relies on platform self-reported demographics (census of their own users) as a proxy.
Full guide
For a complete explanation, read: Panel vs census measurement for CTV: which approach India uses and why it matters