View-through attribution (VTA) for CTV credits a conversion to a CTV ad impression even though no click occurred. The viewer saw the ad on TV, then later purchased on mobile or web. VTA is the primary way CTV campaigns demonstrate conversion impact — because TV is not a clickable medium. But CTV VTA is also the most inflation-prone attribution model in digital advertising. Understanding how it works and where it breaks down is essential for evaluating India CTV performance claims honestly.
What is view-through attribution for CTV
In click-based digital advertising, the attribution chain is: user clicks ad → lands on site → converts. CTV has no click. The attribution chain for CTV is: user sees ad on TV screen → at some point later → user converts on a different device (mobile app, mobile browser, desktop).
View-through attribution bridges the TV impression to the eventual conversion by:
- Recording that device ID X was exposed to the CTV ad at time T
- When a conversion happens on device Y, checking whether device Y is linked to device X (same household or same person)
- If the conversion happens within the VTA window (e.g., 7 days after the impression), crediting the CTV impression with the conversion
How VTA works technically
The technical mechanism depends on how the CTV impression and the conversion are linked:
- GAID-based device matching: If the CTV device (Android TV) has the same GAID as the mobile device used for conversion, the DSP directly matches the impression to the conversion. This is the most accurate path but limited to Android devices where GAID is shared across the ecosystem.
- Household IP matching: The CTV ad impression is associated with a household IP. If a conversion happens from the same IP address within the VTA window, it's attributed to the CTV impression. This is the most common method — and the most prone to inflation. Any conversion from the household (any person, any device) may be credited to the CTV ad.
- Cross-device graph (probabilistic): Attribution vendors (IronSource, AppsFlyer, Branch) use probabilistic models to link CTV device IDs to mobile IDs based on shared signals (IP, location, time). Match accuracy varies widely — typically 50–75% confidence on India household graphs.
Why CTV view-through attribution inflates conversions
Several structural factors cause CTV VTA to over-count conversions:
- Household IP matching attributes to all household converters: A household of 4 people all converting online within the same VTA window after a CTV ad exposure will all be attributed to the CTV campaign — even if only one family member saw the ad, or even if the ad had no influence on any conversion.
- Long VTA windows amplify coincidental correlation: A 30-day VTA window for a CTV campaign running across millions of impressions will capture many conversions that would have happened anyway. In categories with high organic purchase intent (FMCG, mobile phones, OTT subscriptions), the coincidental conversion rate is high.
- No counter-factual baseline: VTA doesn't compare conversion rates of exposed vs unexposed households. It just credits conversions to exposed households. Without an unexposed control group, there's no way to separate CTV-driven conversions from organic ones.
- Co-viewing inflates reach, making VTA credit even broader: If 3 people co-viewed the ad but only 1 device was tracked, the household's conversions from all 3 people are attributed to that single device ID impression.
Appropriate VTA windows for India CTV
| Campaign objective | Recommended VTA window | Rationale |
|---|---|---|
| App installs (e-commerce, fintech) | 1–3 days | Impulse actions happen quickly; longer window credits unrelated installs |
| OTT subscription sign-ups | 3–7 days | Viewer considers, searches, subscribes within a week of exposure |
| Consumer durables | 14–30 days | Research-to-purchase cycle is longer; 30 days reasonable for TVs/appliances |
| Automotive (lead generation) | 14–30 days | Long purchase journey; 30 days captures influenced test-drive inquiries |
| Brand awareness | N/A — use brand lift, not VTA | VTA is a conversion metric; brand lift studies are better for awareness goals |
How to use VTA responsibly in India CTV reporting
Four practices that make CTV VTA more meaningful:
- Use a holdout group: Run the campaign to 90% of eligible devices; hold 10% unexposed. Compare conversion rates between exposed and unexposed groups. The incremental conversion rate (exposed minus unexposed) is the true CTV attribution signal.
- Use short VTA windows by default: Start with 7-day maximum. Compare 7-day vs 30-day attributed conversions — if the 30-day window doubles the conversion count without a logical explanation, the longer window is capturing coincidental conversions.
- Segment VTA by method: Report GAID-matched VTA separately from IP-matched VTA. GAID-matched is more reliable. If IP-matched VTA is >3× GAID-matched, the IP method is likely inflating.
- Triangulate with brand search lift: If CTV campaigns are working, branded search volume should increase during and after the flight. Rising branded search during a CTV flight that corresponds to VTA attribution is a signal of genuine impact.
View-through attribution is a directional signal, not a precise measurement of CTV impact. Treat it as one input among several — alongside brand lift, search lift, and incrementality testing — rather than as the sole performance measure for India CTV campaigns.