If a campaign runs on JioHotstar (reaching 1.5 million devices) and Zee5 (reaching 800,000 devices) simultaneously, the combined reach is not 2.3 million. Some viewers watch both platforms. Adding publisher-reported unique device numbers without deduplicating for cross-platform overlap produces inflated total reach. In India CTV today, there is no easy, independent cross-publisher deduplication tool. Understanding how large the overlap problem is — and what to do about it — is essential for any multi-publisher India CTV campaign.
The cross-publisher overlap problem
Cross-publisher deduplication requires that both publishers share a common identifier (like a GAID or hashed email) so overlapping devices can be identified. In practice, publisher data is siloed:
- JioHotstar uses Jio account data and GAIDs. They don't share device ID lists with Zee5 or SonyLIV.
- Zee5 and SonyLIV use their own registered user systems and GAIDs. No cross-publisher identity linkage.
- DSPs could theoretically track cross-publisher reach if the same DSP was used for all buys, but most India CTV deals are direct IO or publisher-managed, not DSP-managed.
Even when the same DSP is used for multiple publisher PMPs, the DSP can only deduplicate publishers whose inventory comes through the DSP. Direct IO campaigns (bought directly with the publisher's sales team) are invisible to the DSP deduplication.
Estimated overlap rates for India CTV publishers
The CTV audience in India is relatively small and concentrated in affluent urban households — which means platform overlap is higher than it might appear from subscriber numbers alone.
| Publisher pair | Estimated audience overlap | Reason |
|---|---|---|
| JioHotstar ↔ YouTube CTV | 50–65% | Both are default apps on Android TV; heavy JioHotstar users also use YouTube |
| JioHotstar ↔ Netflix (TV) | 30–45% | Netflix skews premium — subset of JioHotstar subscribers also have Netflix |
| JioHotstar ↔ SonyLIV | 25–40% | Sports viewers who watch cricket on both platforms (when rights differ) |
| JioHotstar ↔ Zee5 | 20–35% | GEC viewers with broad OTT subscriptions |
| SonyLIV ↔ Zee5 | 20–30% | Premium OTT subscribers; drama viewers |
| Zee5 ↔ YouTube CTV | 40–55% | Both accessible on Android TV; Zee5 viewers also use YouTube for free content |
These estimates are based on reported multi-platform subscription survey data from MPA (Media Partners Asia) India OTT tracker and FICCI-EY media reports. They are indicative, not verified device-level overlap measurements.
Three approaches to deduplication
1. DSP-managed frequency and reach (partial): Running all publishers through a single DSP (DV360 or TTD) allows the DSP to apply cross-publisher frequency caps using the common device ID. This works only for programmatic buys — not direct IO. DV360's reach planner can model estimated reach with overlap correction for YouTube + Display & Video 360 managed publishers. Coverage is partial for India CTV.
2. Data clean room (for large advertisers): A data clean room allows two publishers to compute overlapping device IDs without either party sharing their raw data. The advertiser, Publisher A, and Publisher B each contribute their device ID lists. The clean room computes the intersection (overlap) and unique reach without exposing individual IDs. In India, this approach is available but operationally complex — it requires clean room agreements with each publisher, and India's largest publisher (JioHotstar) has not made a standardised clean room programme publicly available as of 2026.
3. Post-campaign third-party modelling: Vendors like Nielsen, Kantar, and DoubleVerify can, in some cases, model cross-platform reach deduplication using their own panel and verification data. This provides an estimate, not an exact figure, but is more actionable than simple publisher-reported addition. Nielsen One Ads is expanding coverage in India and may provide CTV-specific deduplication reporting in 2026–2027.
Practical planning without deduplication
For most India CTV campaigns today, true cross-publisher deduplication isn't available. Practical workarounds:
- Apply an overlap discount factor: When summing publisher unique device numbers, subtract an estimated overlap. For two publishers targeting similar SEC A/B urban audiences, a 25–35% overlap discount is realistic. For three publishers, increase the discount to 40–50%.
- Report unduplicated reach as a range: Rather than a single number, report reach as "X to Y million devices (estimated)" with the variance reflecting overlap uncertainty. This is more honest than a false-precision single figure.
- Use DSP-reported unique reach for programmatic portion: The share of the campaign managed programmatically through a single DSP can be reported with DSP-level deduplication. Supplement with publisher-reported figures for direct IO components.
- Sequence, don't overlap: If budget allows, run publishers in sequence (week 1 JioHotstar, week 2 Zee5) rather than simultaneously. Reduces frequency doubling and makes overlap less of a planning problem — though it extends campaign flight length.
When deduplication matters most
Cross-publisher deduplication matters more for:
- Brand awareness campaigns where incremental reach (unique new viewers) is the primary KPI
- Frequency-capped campaigns where over-delivery to the same viewer wastes budget
- High-investment multi-publisher buys (₹50 lakh+) where reporting accuracy is needed for ROI justification
It matters less for campaigns where the goal is saturation reach (be seen by the highest-engagement viewers multiple times) or where the target audience is so niche that even with overlap, total reach is the primary constraint.