Pixel-based attribution for CTV campaigns in India

Pixel-based attribution is the most practical and widely available performance measurement approach for India CTV campaigns. It works by recording a CTV ad impression against a device or household identity, then crediting a downstream action — a website visit, form submission, app install, or purchase — that occurs within a defined time window after the exposure. In a market without a unified cross-screen audience currency, pixel attribution is often the strongest evidence of CTV campaign impact that an advertiser can generate without commissioning a custom brand lift study.

Why pixel attribution for CTV

CTV creates an attribution challenge that doesn't exist for click-based digital media. A viewer watches a 30-second ad on their smart TV. They cannot click. They may pick up their phone an hour later and visit the advertiser's website. The connection between the ad exposure and the site visit is real but invisible to standard last-click attribution — the website visit would appear as direct traffic or organic search, and the CTV campaign would get zero credit.

Pixel-based attribution solves this by logging the CTV impression at the household or device level and then attempting to match it to subsequent web or app sessions from the same household. This "view-through" attribution model is the standard for CTV and for connected TV it provides the closest available signal to causal campaign impact.

How CTV pixel attribution works

The mechanics differ by DSP but the general flow is:

Impression logging: When a CTV ad is served, the DSP (DV360 or TTD) logs the impression with available identity signals — IP address of the household, device IFA (if the CTV device passes one), and timestamp.
Pixel firing on conversion: The advertiser's website or app has a DSP pixel or SDK installed. When a user visits the site or performs an in-app action, the pixel fires and sends the event to the DSP with the user's IP address and other available identifiers.
Matching: The DSP's attribution engine attempts to match the conversion event to a prior CTV impression from the same household (typically via shared home IP address). If a match is found within the attribution window, the CTV impression is credited with the conversion.
Reporting: View-through conversions (VTCs) appear in the DSP's conversion reporting alongside click-through conversions (CTCs) from performance media.

The match key in most India CTV attribution is the home IP address. A smart TV's CTV impression is logged at the household IP. When a mobile phone on the same home WiFi visits the advertiser's site, both events share the same IP — the attribution system connects them. This is called household-level or IP-based attribution.

View-through attribution windows

The attribution window defines how long after a CTV impression a conversion can still be credited to that impression. Standard windows:

1-day view-through: Most conservative. Only credits conversions that happen within 24 hours of seeing the ad. Appropriate for direct-response categories (e-commerce, app install campaigns) where the intent is immediate.
7-day view-through: The most common default for CTV. Balances capturing delayed conversions (someone who saw a home loan ad on Tuesday and visited the bank's website on Saturday) against noise from unrelated conversions that happen to fall in the window.
14-day or 30-day view-through: Used for high-consideration categories (auto, real estate, B2B) where the purchase cycle is long and a CTV ad may plant a seed that converts weeks later. Longer windows capture more conversions but also more false positives.

DV360 default view-through window for CTV is 7 days; TTD default is also 7 days. Both allow the advertiser to customise the window per campaign or advertiser. Set the window to match your category's typical consideration cycle — not the longest available window, which inflates conversion counts without reflecting real attribution.

Site visit uplift measurement

Beyond raw view-through conversions, the most defensible CTV attribution metric for brand campaigns is site visit uplift: the incremental increase in website visits attributable to the CTV campaign. This is measured by comparing site traffic during the campaign flight against a matched baseline (same period prior year, or a geographic holdout region that didn't receive the campaign).

Site visit uplift is particularly useful in India because:

It is measurable with Google Analytics or any standard web analytics tool — no additional cost or vendor relationship required
It captures the full household-level effect, not just the single device that was cookied
It is harder to inflate than raw view-through conversions (which can be gamed by setting long windows)

A clean site visit uplift methodology for India CTV: run DV360 campaigns geo-targeted to specific cities (Delhi, Mumbai, Bangalore). Use cities of similar size and profile that didn't receive the campaign as a holdout. Compare indexed site visits from target cities vs holdout cities during and after the campaign flight. Incremental visits from target cities are attributable to the CTV campaign with reasonable confidence.

App install attribution via MMPs

For advertisers with mobile apps, Mobile Measurement Partners (MMPs) — AppsFlyer, Adjust, and Branch — provide CTV-to-app attribution. The MMP SDK in the app captures install and in-app event data. When a new install occurs, the MMP checks whether the installing device's IP matches a household that received a CTV impression from the campaign (within the attribution window).

AppsFlyer and Adjust both support CTV view-through attribution in India. The integration requires:

MMP SDK in the advertiser's app
CTV impression data shared from the DSP to the MMP (via postback or data export)
Attribution window configuration in the MMP dashboard

App install attribution via MMPs is the most reliable CTV attribution approach for performance advertisers because it uses deterministic install events rather than probabilistic site visit matching. It is standard practice for India fintech, e-commerce, and gaming advertisers running CTV alongside performance media.

India CTV specifics

A few India-specific factors affect pixel attribution quality:

Mobile hotspot connectivity: Smart TVs connecting via Jio 4G/5G mobile hotspot share a carrier-grade NAT IP with thousands of other mobile users — not a stable home IP. IP-based household matching fails for this segment. Attribution for mobile-broadband-connected CTV households in India is materially weaker than for JioFiber or Airtel Fiber households with stable residential IPs.
Dynamic IPs: Even JioFiber and Airtel Fiber IPs rotate periodically. IP-based attribution works best within a 24–72 hour window before IP rotation degrades match accuracy. This is one reason why shorter view-through windows (1–7 days) produce cleaner attribution data in India than longer windows.
Shared IPs in dense urban housing: In apartment complexes with shared building WiFi or router-based connectivity, multiple unrelated households may share the same IP. Attribution from these environments will overcount conversions. Metro areas in India with dense apartment stock (Mumbai, Bangalore, Hyderabad) are particularly affected.
DPDPA compliance: Any pixel that collects user-level data in India must comply with the Digital Personal Data Protection Act. Consent should be obtained before firing conversion pixels on Indian users. DSP pixels with consent management integration (Google Consent Mode, TCF) are the compliant approach.

Limitations and what to watch for

Correlation vs causation: A household that saw a CTV ad and visited the website may have been going to visit anyway. View-through attribution overcounts when the baseline visit rate is not controlled for. Use uplift methodology rather than raw VTC counts where possible.
Window gaming: Setting a 30-day view-through window makes conversion numbers look impressive but most of those conversions would have happened without the ad. Always report the window alongside conversion numbers.
Cross-publisher deduplication: If a campaign runs on both JioHotstar (DV360) and SonyLIV (TTD), both DSPs will claim view-through credit for the same conversion if the household IP matched in both systems. Deduplicate across DSPs using a single MMP or pixel as the source of truth.