Ad-Supported Video on Demand (AVOD)

MediaRadar’s AVOD measurement covers ads served on ad supported platforms streaming on-demand programming. Collection does not include Live TV content.

MediaRadar has twelve platforms in its historical data library:

MediaRadar regularly reviews coverage and will add relevant platforms as feasible.

MediaRadar's AVOD ad data is captured through panel-based collection partnerships that passively record advertising served to real consumers across leading streaming services. As of January 1st, 2026, the AVOD dataset is a unified view drawn from two complementary collection sources that operate in parallel.

The combined panel uses software installed on panelists’ in-home devices and opt-in apps installed directly on Smart TVs and other connected devices to capture ads as they are served during streaming sessions. Panel collection has been in place since January 2023, replacing MediaRadar’s prior crawler-based approach.

Together, the two panel-based collection sources capture in-stream video ads across leading AVOD services on Smart TV, mobile, and desktop devices. Both rely on actual viewing behavior, providing substantially broader advertiser representation than the synthetic behavioral profiles used by MediaRadar’s legacy crawler approach.

The combined panel base is geographically distributed,

optimizing collection of national, regional, and local advertisers.

MediaRadar does not directly manage panel recruitment for either source. Both collection sources rely on anonymized, opt-in panels and do not collect personally identifiable information (PII). Because of this, demographic information on individual panelists is not available.

For each ad occurrence, metadata is reported by date, platform, advertiser, creative, and commercial length. At this time, program and daypart details are not harvested.

Spending estimations are determined using a statistical model, taking into account the platform’s valuation of advertising and the weighted sample of observational data collected across the platform. Streaming ads are collected using a sample-based methodology as a census of all ads served to the OTT ecosystem is not available.

The data model takes into account the platform’s advertising valuation based on a number of variables, including company financials; subscriber thresholds along with average ad revenue per subscriber; third-party benchmarks as sourced through eMarketer, IAB, and others; engagement metrics as sourced through SimilarWeb; time spent with video as sourced through third party surveys; ad load reports as sourced from MediaRadar’s systems; and industry centric CPMs as sourced from Guideline/SMI/SQAD. Spending estimations are an extrapolation of investment served across all devices for the platform. The data model weights the observational data to derive Advertiser spending estimates.

Ad expenditures are an estimation of the relative, comparative value of the advertisers competing in the OTT landscape, and spend estimations are not tailored to the brand’s specific buy parameters. Additionally, spend estimations reflect a full aggregation of delivery across devices and screens, in this way, there is not variable treatment or reporting by device type.
​

MediaRadar estimates AVOD ad impressions at the platform level and allocates them to advertisers. For each platform, a monthly total of ad impressions is established by the same data model used for expenditure estimation, drawing on subscriber and engagement metrics, time spent with video, ad load reports, and third-party benchmarks. The platform total is distributed across the advertisers, brands, and products observed on the platform in proportion to their share of collected ad occurrences. Impressions and expenditures are consistent with one another: reported expenditure reflects the platform CPM applied to reported impressions.

Expenditure = (CPM) * (Impressions / 1000)

Impressions are available from the start of collection for each platform (see Coverage); the length of impression history therefore varies by platform. As with expenditures, impressions are an estimate of relative delivery based on a sample of observed occurrences and reflect a full aggregation across devices and screens.
​

Prior to the release of final monthly ad expenditures and impressions, preliminary estimates are updated daily and follow the same methodology as final estimates, without the calibrations applied at month close. Each daily update uses all ad occurrences collected since the beginning of the month and restates the month to date, so estimates refine as more advertisers and creatives enter the sample.

Preliminary impressions are allocated from the platform’s impressions in the most current month of final data, in proportion to each advertiser’s share of collected occurrences. To limit early-month volatility, an advertiser’s preliminary average daily impressions on a platform are capped at its average daily impressions in the most current month of final data; advertisers not present in that month are reported as observed. Preliminary expenditures are calculated by applying the platform CPM to preliminary impressions. For the most recent days not yet processed, an advanced estimate based on the previous month’s average daily final spend serves as a placeholder until it is replaced by the refined estimate.

Preliminary estimates are available approximately two days after the day of collection, are directional in nature, and are replaced by final estimates when the month is finalized, approximately 3–4 weeks after month close. Brands reported in preliminary estimates reflect confirmed sightings of ads served on the platform.
​

Each collected ad is matched against a master library of previously detected AVOD video ads. If a match is found and confirmed as advertising material, a record is entered into the occurrence file indicating the platform and date of collection, and the associated details housed in the video library (e.g., product name, brand, advertiser and category).

If a match is not found in the master library, the ad is sent to an editor for manual review. Once identified as either a valid ad or a non-ad, the signature is sent to the master library for permanent storage and future use in the matching process. For a new ad, the editor classifies it into our proprietary central database of advertised products. A record is entered into the occurrence file indicating the platform and date of collection, and the associated details housed in the video library (e.g., product name, brand, advertiser and category).

To access our full 360 Methodology, log in to the 360 platform and click on the Help Icon located at the bottom-right corner of the application.

​