Audience Measurement

Unique reach in CTV: how to measure how many people actually saw your ad

Unique reach is the number of distinct people, households, or devices that were exposed to your advertising at least once during a campaign flight. It is different from total impressions, which count every ad play including multiple views by the same person. If a household sees your ad 10 times across a campaign, they contribute 10 impressions but only 1 unique reach. Understanding and tracking unique reach is fundamental for evaluating how broadly a campaign penetrated your target audience — and for managing frequency to avoid over-exposure.

Unique reach vs total impressions: why the distinction matters

Two campaigns can have identical impression counts but very different reach profiles:

  • Campaign A: 10 million impressions delivered to 10 million unique devices — average frequency of 1
  • Campaign B: 10 million impressions delivered to 1 million unique devices — average frequency of 10

For a brand-building campaign trying to drive awareness among as wide an audience as possible, Campaign A is more valuable. For a retargeting or performance campaign where repeated exposure is used to drive conversion, Campaign B might be intentional. But you cannot distinguish these scenarios from impressions alone — you need unique reach.

This is why unique reach and average frequency are always paired: Unique Reach × Average Frequency = Total Impressions. Any two of these figures give you the third.

Device, household, and person-level reach

Unique reach is measured at three levels, and the number changes significantly depending on which level you use:

Device-level unique reach

The most common level for CTV measurement. Each CTV device has a unique identifier — a device ID or advertising ID. The platform counts how many distinct device IDs received at least one impression. This is straightforward to measure and is the default for most platform-reported CTV campaign data.

The limitation: a single smart TV in a household counts as one device even if four people watch it. And the same person using three different devices (smart TV, phone, laptop) counts as three devices. Device-level reach neither correctly accounts for co-viewing nor cross-device behaviour.

Household-level unique reach

Households are typically identified by IP address — devices sharing a home network share an IP. Household-level reach counts distinct IP addresses (or household IDs when available) exposed to the campaign. This is a more stable unit for CTV measurement than device IDs, since a household is typically a consistent economic unit relevant to most advertiser objectives.

Household-level reach is still an undercount of persons reached (because of co-viewing) and an overcount of individuals (because shared devices are grouped together).

Person-level unique reach

The most meaningful for most advertising objectives — how many individual people saw the ad? Person-level reach requires either deterministic identity (a logged-in account linked to a named person) or a panel-calibrated estimate that adjusts device and household data based on demographic composition and co-viewing factors.

Person-level reach is the gold standard and the hardest to measure accurately in CTV. Most India CTV campaigns report device or household-level reach, not true person-level reach.

How frequency-capping requires unique reach data

Frequency caps — the maximum number of times a device or user can see an ad within a given period — only function correctly when the ad system can identify unique viewers. The mechanism:

  1. A device makes an ad request (a CTV app asks for an ad to play)
  2. The ad server checks the device's impression counter in its frequency database
  3. If the device has already received the maximum allowed impressions, the ad server suppresses this device and serves a different ad or no ad
  4. If under the cap, the impression is delivered and the counter is incremented

This process is straightforward within a single platform. When a campaign runs across multiple platforms (JioCinema + SonyLIV + YouTube), frequency caps cannot be enforced across platforms unless those platforms share a common user identity. Without cross-platform identity, you can cap frequency at 3 on each platform — but the same household could receive 3 impressions from each of 3 platforms = 9 total, with no single platform aware of the over-exposure.

Which India platforms provide user-level reach data

The availability and quality of unique reach data varies considerably across India's CTV ecosystem:

  • JioHotstar (JioCinema + Hotstar post-merger): Has strong logged-in user data from Jio's subscriber base. Can provide user-level unique reach for logged-in inventory. Guest viewing generates device-level data only.
  • SonyLIV: Account-based streaming. Unique reach reporting is available at the subscriber account level for logged-in viewers. Reporting detail shared with advertisers varies by deal type.
  • Zee5: Account-based. Similar position to SonyLIV. Campaign reports include unique device and user reach counts.
  • YouTube CTV: Google provides unique reach data within its own ecosystem across YouTube surfaces (CTV, mobile, desktop). This is among the stronger unique reach reporting available in India given Google's identity infrastructure.
  • Samsung Ads / OEM advertising: Delivers at the device level (TV IDs). Household-level reach is available via IP clustering. Person-level data is not natively available but can be estimated with panel calibration.

Platforms that do not require login (free ad-supported services without mandatory accounts) can only report device-level reach. The share of impressions from logged-in vs guest views affects the quality of unique reach data significantly.

Practical guidance for tracking unique reach in India CTV

  • Always request unique device reach, not just impressions: Any platform should be able to provide unique device reach as a campaign metric. If a platform only reports impressions, push for the unique device breakdown.
  • Ask about the percentage of logged-in vs guest inventory: This tells you what share of the unique reach figure is user-level (more accurate) vs device-level (less accurate).
  • Apply a co-viewing multiplier for TV-screen inventory: Unique device reach on a CTV device likely represents more than one person. Apply a co-viewing factor (directionally 2–3x for India prime-time) to estimate unique person reach.
  • Do not sum reach across platforms: Adding JioCinema unique reach + SonyLIV unique reach + YouTube unique reach does not give you campaign total unique reach — it gives you a duplicated figure. Report each platform separately or use a methodology note when combining.
  • Track reach curves over time: A reach curve shows how unique reach accumulates with each additional impression served. Reach curves are useful for identifying diminishing returns — when additional impressions are primarily generating frequency rather than new unique reach.