Google announced on May 27, 2026, that the Google Ads API will adopt a "Total Co-view" definition for reach metrics starting June 2, 2026. The change - published on the Ads Developer Blog by Mattia Tommasone of Google's Advertising and Measurement Developer Relations team - affects seven campaign-level metrics and requires no code changes from developers. Reported values for reach and frequency will shift from the launch date forward.

What changes on June 2

The update moves the definition of video view metrics away from a device-centric model toward one that counts every person present in front of a connected TV screen at the moment an ad plays. According to the Google Ads Developer Blog, the updated definition "will include all individuals who viewed the ad on connected TV devices, including cases where multiple people are watching YouTube together on the same TV screen."

Seven campaign-level metrics are affected by the switch:

  • metrics.unique_users
  • metrics.average_impression_frequency_per_user
  • metrics.unique_users_two_plus
  • metrics.unique_users_three_plus
  • metrics.unique_users_four_plus
  • metrics.unique_users_five_plus
  • metrics.unique_users_ten_plus

All seven sit at the campaign level and collectively form the primary set of reach and frequency fields available through the API. The change is classified as a data definition change rather than a structural API modification. According to Google, "no API version migration or code changes are required to see this new behavior, as it is a data definition change that will be applied automatically." That means development teams do not need to update client libraries, regenerate service stubs, or modify query structures. The numbers will simply change.

What the threshold metrics represent

Understanding the affected metrics matters for interpreting how numbers will shift. metrics.unique_users counts distinct individuals who saw an ad at least once. metrics.average_impression_frequency_per_user divides total impressions by that unique user count - a standard frequency calculation. The remaining five fields are frequency distribution thresholds: metrics.unique_users_two_plus counts how many unique users received at least two exposures, and the series continues through the five-plus and ten-plus thresholds.

These distribution metrics are especially important for campaigns that apply frequency management. An advertiser running a reach objective campaign may set a frequency cap of three exposures per person per week. The metrics.unique_users_three_plus field shows how much of the audience has hit or exceeded that boundary. If co-viewing inflates the unique user pool - because additional people in the same room are now counted - the same absolute number of impressions will distribute across a larger denominator, and the frequency distribution will shift accordingly.

How co-viewing is measured

Co-viewing occurs when multiple people watch YouTube on a connected TV (CTV) device simultaneously and see an ad at the same time. The concept is not new to television measurement. Broadcast ratings have long credited every person in the household who was present during an ad break, not just the individual who pressed the power button on the remote. What makes it technically complex for digital measurement is that a CTV device typically registers as a single account or device ID - all additional viewers around that screen are invisible to device-level tracking alone.

Google addresses this using two overlapping data sources. According to the Google Ads Help Center co-viewing definition page, the measurement combines "Google's connect TV panel data and real-time, census-level surveys to assess co-viewing behavior across 100+ countries via 70+ languages worldwide." Statistical methods are applied to eliminate bias and ensure the results accurately represent YouTube's viewing population rather than reflecting artefacts of the panel composition.

The dynamic reach model

The resulting data does not produce a single global co-viewing multiplier. Instead, it feeds a dynamic reach model. According to the Help Center documentation, the model predicts co-viewing behavior based on four variables: the day-of-week, time-of-day, country, and user demographics. This is a meaningful technical distinction from applying a flat adjustment factor. A Saturday evening broadcast in South Korea will carry different household viewing patterns than a Tuesday midday stream in Sweden, and the model is designed to reflect those differences rather than averaging them away.

This methodology has been part of Google's planning layer for some time. In April 2022, Google introduced co-viewing metrics across Google Ads and Display & Video 360, making them available in planning and measurement tools by the end of Q2 2022. At that point, a Nielsen study cited by Google found that 26% of the time, multiple adults aged 18 and over were watching YouTube together on a TV screen, compared to 22% on linear TV. At the same time, Google announced YouTube CTV and YouTube TV would be available in Nielsen Total Ad Ratings, with co-viewing included in Nielsen DAR guarantees in the US. The June 2026 change embeds co-viewing into the standard metric definitions rather than treating it as an optional enrichment layer.

The API had already taken smaller steps in this direction. The Google Ads API v19.1 release in April 2025 added two separate metrics columns - coviewed_impressions and primary_impressions - along with adjusted demographic segments that provide more accurate breakdowns of co-viewed video views. The Total Co-view switch on June 2 goes further, folding co-viewers directly into the foundational unique user counts rather than reporting them as a supplementary column.

Alignment with offline and third-party measurement standards

Google frames the June 2 transition explicitly as a standards alignment. According to the Google Ads Developer Blog announcement, the goal is to bring Google Ads reporting in line with "offline media and third-party measurement standards, ensuring consistency across all reporting surfaces, including the Google Ads UI and Editor."

That consistency argument addresses a real structural problem for media planners. Television audience measurement - whether through Nielsen, Kantar, or national measurement bodies - credits every person present in the room during an ad break. When a YouTube CTV campaign was measured using a device-centric model, the resulting reach number was fundamentally smaller than an equivalent linear TV number, not because fewer people saw the ad, but because the measurement methodology was narrower. Direct budget comparisons across channels were distorted by that definitional gap.

Google Ads introduced Cross-Media Reach measurement in June 2024, offering deduplicated reach and frequency across video campaigns, with the ability to compare YouTube performance against linear TV in select regions. The Total Co-view update strengthens the foundation beneath that tool: the individual reach figures feeding the cross-media comparison will now reflect co-viewing behavior on the YouTube side, making the comparison less asymmetric.

Third-party measurement providers have been building parallel infrastructure. Nielsen expanded its YouTube CTV ad measurement to Australia in August 2024, incorporating co-viewing data into Nielsen ONE Ads and deduplicating exposures across computer, mobile, and CTV surfaces. AudienceProject expanded its YouTube measurement to new European markets in May 2025, offering reach and frequency measurement across all devices with co-viewing included. Both of these third-party implementations were already accounting for shared viewing. The Total Co-view change brings Google's own first-party API reporting closer to the methodology those vendors had already adopted.

The Help Center documentation notes that Nielsen DAR guarantees remain available. According to the documentation, "advertisers and agencies can continue to buy on Nielsen DAR" - meaning the definition change affects organic API reporting but does not displace existing contractual measurement arrangements.

Impact on reporting and dashboards

The practical preparation required is limited but not trivial. No code changes are needed, but internal reporting systems, dashboards, and budget pacing tools that display metrics.unique_users or any of the six related fields will show different numbers from June 2 onward. The direction of change is predictable: CTV campaigns reaching households where multiple people view the same screen will report higher unique user totals and, depending on the campaign's impression distribution, different frequency profiles.

According to the developer blog, developers and advertisers "should be prepared for a potential change in reported reach and frequency values for dates beginning with the launch date." The recommendation in the post is to update internal documentation or dashboards to reflect the shift in measurement methodology. Historical data before June 2 will remain under the previous definition - there is no retroactive restatement of earlier reporting periods.

That creates a data discontinuity point. Any trend line tracking metrics.unique_users across months will have a structural break at June 2 that is not attributable to campaign performance changes, inventory fluctuations, or audience behavior shifts. Teams running automated anomaly detection or machine learning models on these metrics should flag the date as a methodology boundary.

Several other API changes land in close proximity. In April 2026, Google announced that the API will expand product reporting for Video, Demand Gen, and App campaigns starting June 15, 2026, adding cost and conversion metrics to campaign types that previously returned only impressions and clicks. The data retention change announced in May 2026limits granular reach and frequency metrics - including the unique_users series affected by the co-viewing change - to a three-year retention window, shorter than the 37-month window applied to other metrics. The co-viewing definition switch on June 2 thus arrives in the middle of a compressed period of API reporting changes, all of which affect how reach and frequency data is interpreted.

DV360 updated its frequency cap and reach metrics in August 2025, replacing bid-request-based frequency calculations with total media cost-based calculations to more accurately represent auction dynamics. That update and the Total Co-view change follow the same pattern: replacing technically convenient approximations with more accurate representations of underlying audience behavior.

The broader context of CTV measurement

The significance of this change reaches beyond its immediate technical scope. Connected TV is now YouTube's fastest-growing viewing surface. YouTube's Q1 2026 TV update introduced conversational AI search, chapter navigation, family groups, and other features designed to deepen the platform's integration with living room viewing. YouTube Brandcast 2026 in May 2026 announced two-click CTV checkout via Google Pay and expanded retail data partnerships, extending the transactional layer into connected TV advertising. As the platform's commercial ambitions on television scale up, the measurement infrastructure is expected to match how television audiences have always been counted.

The Help Center co-viewing definition document notes that panels show "multiple people are watching YouTube together on TV screens, a consumer behavior characteristic of linear television viewership as well." A device-only measurement model for CTV does not capture that behavior. When an agency presents a media plan comparing a linear TV schedule against a YouTube CTV buy, the reach and frequency numbers on both sides of the comparison now reflect the same underlying concept - household presence - rather than linear TV counting the room and YouTube counting the account.

The Google Ads cross-campaign brand metrics system introduced in December 2024 already incorporated co-viewing data into its underlying reach model, having updated the model in April 2022 when co-viewing was first introduced. The June 2 change propagates that co-view adjustment into the core API metrics fields that third-party tools and internal reporting systems consume directly.

The Bid Manager API expanded reach data inclusion in November 2024 by introducing a confidence-based filter that allowed developers to include data points previously excluded due to lower statistical significance. That change and the Total Co-view switch address different dimensions of the same underlying measurement challenge: ensuring that API-level reach data reflects the actual breadth of advertising exposure as accurately as current modeling allows.

Developer support and resources

Questions and concerns about the update can be directed to the "Google Advertising and Measurement Community" Discord server, as specified in the developer blog post. The post was published on May 27, 2026, under the google_ads_apilabel on the Ads Developer Blog, and was authored by Mattia Tommasone of Google's Advertising and Measurement Developer Relations team.

The co-viewing definition page in the Google Ads Help Center provides further documentation on how co-viewing data integrates with the Reach Planner tool, covering reach forecasts, campaign reports, and the relationship between co-viewing metrics and Nielsen DAR guarantees.

Timeline

Summary

Who: Google, through Mattia Tommasone of the Advertising and Measurement Developer Relations team, announced the change via the Google Ads Developer Blog on May 27, 2026. The update affects developers and advertisers using the Google Ads API to measure reach on campaigns that serve on connected TV surfaces.

What: Starting June 2, 2026, seven campaign-level reach metrics in the Google Ads API transition to a "Total Co-view" definition. The change counts all individuals who watched an ad on a connected TV device, including secondary viewers sharing the same screen. The affected fields are metrics.unique_users, metrics.average_impression_frequency_per_user, metrics.unique_users_two_plus, metrics.unique_users_three_plus, metrics.unique_users_four_plus, metrics.unique_users_five_plus, and metrics.unique_users_ten_plus. No API version migration or code changes are required.

When: The announcement was published on May 27, 2026. The new definition takes effect for all reporting data from June 2, 2026. Historical data before June 2 remains under the previous measurement definition and will not be restated.

Where: The change applies across all Google Ads reporting surfaces that surface these metrics - the API, the Google Ads UI, and Google Ads Editor. The underlying co-viewing methodology covers more than 100 countries and more than 70 languages worldwide.

Why: The update brings Google Ads reporting into alignment with offline media and third-party measurement standards that have historically counted all viewers present during an ad exposure, not just the account holder associated with the device. The change enables more direct comparisons between YouTube CTV campaigns and linear TV buys, and reflects how shared viewing on television screens works in practice.