The Video Advertising Bureau this week distributed a guide walking marketers through five methods of gathering television viewership data, cataloguing the limitations built into each one, from opt-in requirements on smart TVs to income bias in set-top box households.

The document, titled "What's the Deal with Viewership Data Collection?", is the first installment in a five-part educational series the trade organization calls "What's the Deal with," according to VAB. The series aims to walk marketers through five areas of measurement over the course of 2026: viewership data collection, identity, engagement, outcomes, and what the organization terms "what's next." Two later entries in the series, covering engagement and identity, have already reached VAB members and qualified marketers, arriving on July 30 and July 29, 2026, respectively, even though this viewership data collection guide addresses the first of the five topics chronologically in the framework.

According to VAB, viewership data collection is the systematic process of gathering and measuring information from different sources to be used for decision-making. That definition sits at the center of a document built around six questions the organization says it hears repeatedly from marketers: how viewership gets measured, what census-level data means for advertising, whether a full cross-platform picture is achievable, how targeting can move beyond demographics, how measurement can be future-proofed, and what role panels still play in an industry increasingly reliant on large-scale device data.

Five methods, five sets of limitations

The guide breaks viewership measurement into five collection types, each defined by where the data originates and what it can and cannot show. Automatic content recognition, run on internet-connected smart TVs, identifies advertising and programming through audio or video fingerprints. According to VAB, this method can capture both linear and over-the-top viewing, but it depends on a consumer opt-in: households are required to activate the feature before a set can begin collecting viewing data. Companies cited as examples in the guide include Vizio Ads and LG Ad Solutions.

Set-top box measurement, drawing on what the industry calls return path data, works differently. Calibrated boxes installed in subscriber homes generate household-level viewing information, giving what VAB describes as more granular data across a large geographic footprint. The trade-off, according to VAB, is representational: set-top box households tend to skew toward higher incomes, a pattern that can introduce bias into the resulting dataset. Spectrum Reach and Comcast Advertising appear as example providers for this collection type.

Panels remain the third method, collecting viewing habits, behaviors, and demographic information from a small group of people intended to represent a larger population. When panels are stable, VAB states, they can provide a common view of the television landscape that other methods struggle to replicate. Their central limitation is cost: panels are expensive to maintain, and they can run into representation challenges as viewing habits fragment across a growing number of platforms. Nielsen and TVision are listed as example companies operating in this category.

Streaming measurement, the fourth type, covers over-the-top services delivering content from broadcast and cable networks alongside dedicated streaming providers such as Sling TV, YouTube TV, HBO Max, Netflix, Peacock, Hulu, and fuboTV. According to VAB, data collected through streaming can arrive in real time, which creates flexibility for campaign reporting and optimization. The document notes this method is more likely to be restrictive in how it grants access to measurement, attribution, and reporting data, a limitation that affects how buyers can act on what they see. Peacock and Roku are named as examples.

The fifth and broadest category is big data, which VAB defines as census-level information comprised of multiple sources, potentially including automatic content recognition, set-top box, panel, and streaming inputs combined. More aggregated data, the guide argues, can mean greater opportunity to surface valuable insights. But managing datasets at that scale carries its own cost: according to VAB, large datasets can become unwieldy, requiring substantial processing before the information becomes actionable. Comscore and iSpot.tv are cited as example providers.

What census-level data changes

A recurring theme throughout the guide is the distinction between sample-based and census-level measurement. VAB defines census-level data as a state in which every viewer within an ad environment universe is counted, rather than estimated from a smaller representative sample. The document lists five ways this shift can help marketers: reducing ad waste, enabling more representative measurement, identifying audiences with greater granularity, extending coverage across people, platforms, and geographies, and attributing response more precisely.

The guide organizes the argument for census-level data along three axes. On the question of who is measured, VAB states that census-level data allows for more demographically representative measurement, adding that this creates economic growth opportunities for marketers targeting audiences that have historically not been counted fairly. On the question of what is measured, the document points to consumers who have broadened their own definition of television, following premium content across an expanding set of platforms and devices; counting viewership only on one screen, in this reading, leaves gaps. On the question of where measurement occurs, VAB argues that census-level data creates flexibility to measure more precisely at local, regional, and national levels simultaneously.

Cross-platform measurement and five baseline metrics

The guide addresses directly whether a marketer can obtain a full picture of viewership across a cross-platform video campaign. According to VAB, advancements in data collection have enabled measurement providers to deliver that full picture, producing what the document calls a plethora of ways to evaluate campaign success. Even so, VAB narrows the field to five metrics it argues should be treated as table stakes for any cross-platform evaluation: total reach, defined as campaign reach across all platforms; total impressions, meaning the total number of exposures delivered by a campaign; average frequency, the average number of times a given viewer has encountered the campaign; unique reach, split between the reach of linear-only viewing and the reach of digital-only viewing; and deduplicated reach, which identifies who saw a campaign on both platform types against who saw it on only one.

Beyond the standard demographic breakdown, the guide identifies five additional ways marketers can target and measure impact: geographic targeting at local, regional, and national levels; psychographic targeting based on attitudes, feelings, interests, values, opinions, and lifestyle; behavioral targeting tied to actions consumers take, such as shopping for a product or qualifying as a light or heavy viewer; device-based targeting tied to the specific hardware consumers use; and first-party data collected directly by a marketer from its own customers. According to VAB, these capabilities, layered onto an audience-based buying approach, can help marketers reach the right audience, heighten engagement, and drive outcomes across the full marketing funnel.

Three types of panels, and why panels are not disappearing

Despite the industry's shift toward big data, the guide devotes its final question to panels specifically, outlining three distinct types that continue to play a role in measurement. A measurement panel collects viewing habits, attention-related behaviors, and demographic information from a small group meant to represent a larger population; its chief limitation, according to VAB, is that it can lack representation of certain household types, such as over-the-air-only homes. A calibration panel is used alongside another data asset, such as big data, and analyzed to locate gaps or biases within that larger dataset; the risk here, the guide notes, is that any bias present in the panel itself can be passed along into the dataset it is meant to correct. A validation panel exists to demonstrate mathematically that measurement gaps, whether representational or otherwise, have actually been addressed; VAB describes this as requiring a well-trained machine learning model to function.

According to VAB, the future of television measurement will likely depend on both big data and panels working together, with panels serving as a complement rather than a replacement. In many instances, the guide states, big data would serve as the primary source of viewing information, while panels calibrate and validate that data to ensure its accuracy. The document frames this as a vital step: big data used without calibration and validation, in this framing, is data whose accuracy cannot be confirmed.

Best practices and a broader push

The guide closes with four recommendations addressed to marketers. According to VAB, marketers should know where their viewership data originates, request that cross-platform viewership metrics be treated as table stakes within their own reporting, recognize that television has become a data-driven medium capable of powering customer targeting, acquisition, and engagement, and prioritize forward planning to position themselves for measurement success as the landscape continues to shift.

This is not VAB's only measurement initiative underway in 2026. The trade organization separately maintains an Advanced Measurement Solutions Directory, a resource PPC Land covered after it expanded to 20 vendors on July 21, 2026, organized around the same four measurement categories that anchor the "What's the Deal with" series: viewership data collection, identity, engagement, and outcomes. That directory update landed at a moment PPC Land's earlier reporting described as one in which the television industry was questioning who is entitled to grade its own advertising and which figures can be trusted, a concern sharpened by consolidation among measurement vendors themselves.

Why this matters for marketers

For advertising buyers navigating television and streaming budgets, the distinctions VAB draws between collection methods carry direct planning consequences rather than abstract technical interest. A campaign measured primarily through set-top box data, for instance, risks skewing toward higher-income households by the nature of the underlying sample, a limitation the guide states plainly rather than glossing over. A campaign relying on automatic content recognition depends on how many households in a given market have opted into that feature on their smart TVs, which is not guaranteed to be uniform across demographics or regions.

PPC Land's prior coverage situates these method-level distinctions within a wider industry recalibration already underway. Nielsen's Big Data + Panel methodology, which PPC Land reported became the standard currency for the 2025 Upfront negotiations after merging a 42,000-home panel with device-level inputs from 45 million households and 75 million devices, exemplifies precisely the hybrid approach VAB's guide recommends: big data for scale, panels for calibration. That system has continued to generate friction. PPC Land reported that Nielsen altered seven TV currency metrics scheduled for deployment on August 31, 2026, forcing buyers to recheck their planning assumptions ahead of the fall season, an illustration of how frequently the underlying methodology can shift beneath a currency that advertisers treat as stable.

The Media Rating Council and the Interactive Advertising Bureau finalized attention measurement guidelines in November 2025, a development PPC Land's directory coverage noted established minimum requirements for transparency and comparability after years of inconsistent vendor methodologies. Those guidelines also cautioned that attention should not be treated as a direct measure of campaign outcomes, a distinction relevant to marketers reading VAB's engagement-focused entry in the same series, which PPC Land covered after its July 30, 2026 distribution.

VAB's own directory carries what PPC Land's earlier reporting called an inherent framing worth keeping in view: VAB is a trade organization whose members are premium multiscreen television sellers, giving it a direct interest in findings and vendor rosters that support the value of television advertising broadly. The viewership data collection guide, similarly, is an educational resource distributed to VAB members and qualified marketers rather than an independently audited methodology comparison. It offers a map of collection types and their trade-offs. It does not, by itself, resolve which vendor's specific numbers a buyer should trust for a given campaign.

The census-level data framing running through the guide also connects to a broader industry transition PPC Land has tracked across 2026. Kantar Media's relaunch as Fifty5Blue, reported by PPC Land in February 2026, combined panel data from more than 5,000 homes with census-level online viewing data using a comparable logic to what Nielsen and VAB both describe: neither a sample alone nor an unvalidated mass of device data is sufficient on its own in a fragmented viewing environment spanning linear television, streaming, connected TV, and social platforms.

Timeline

  • January 24, 2025: Nielsen declares the end of its stand-alone panel-based TV ratings system, announcing a transition to a combined big data and panel methodology.
  • September 2, 2025: Nielsen launches Big Data + Panel measurement for the 2025 broadcast season, merging its 42,000-home panel with device-level inputs from 45 million households and 75 million devices.
  • November 2025: The Media Rating Council and the Interactive Advertising Bureau finalize attention measurement guidelines establishing minimum requirements for transparency and comparability.
  • February 25, 2026: Kantar Media exits Kantar Group and relaunches as Fifty5Blue, combining panel data with census-level online viewing data under a similar hybrid logic.
  • July 21, 2026: VAB expands its Advanced Measurement Solutions Directory to 20 vendors, organized around viewership data collection, identity, engagement, and outcomes.
  • July 29, 2026: VAB distributes the third installment in its five-part series, "What's the Deal with Identity?"
  • July 30, 2026: VAB distributes the second installment in the series, "What's the Deal with Engagement?"
  • Today: VAB distributes "What's the Deal with Viewership Data Collection?", the first installment in the series, outlining five collection methods and their respective limitations.

Summary

Who: The Video Advertising Bureau, a trade organization representing premium multiscreen television sellers, published the guide for its members and what it calls qualified marketers.

What: VAB distributed "What's the Deal with Viewership Data Collection?", an educational document outlining five methods of gathering television viewership data, census-level data's role in measurement, five baseline cross-platform metrics, five ways to target audiences beyond demographics, and three types of panels that continue to inform modern measurement.

When: VAB distributed the guide today. It is the first installment chronologically in a five-part series that also includes entries on engagement, distributed July 30, 2026, and identity, distributed July 29, 2026.

Where: The guide addresses measurement across the United States television and streaming ecosystem, spanning linear broadcast, addressable television, connected TV, and cross-platform streaming environments.

Why: Marketers face an expanding set of choices for planning, buying, and measuring video campaigns, and VAB frames the guide as a response to common questions the organization says it hears repeatedly from advertisers navigating that fragmented measurement landscape.