YouTube began adding estimates of people watching together on television screens to its public video counts Nov. 16, 2026, changing the meaning of a familiar audience indicator for brands and publishers. The company’s API revision history says the change covers long-form videos, live streams and Shorts. Live concurrent-viewer counts also incorporate people sharing a TV. YouTube says engaged views, creator earnings, advertising revenue and Partner Program eligibility are unaffected. The estimate changes what the public total represents; it does not establish longer viewing on devices.
Rapid News Brief: Ad billing remains distinct
Google’s advertising measurement guide separates public totals from TrueView views. The latter determine charges for cost-per-view bidding in Video View Campaigns. TrueView counts exclude estimated additional viewers sharing a television: a qualifying ad playback on one device produces one TrueView view.
Public views can include that estimated shared audience, however, and Google says it does not use public views to bill YouTube advertising campaigns. The guide also distinguishes public views from impressions. It says a campaign reaching television screens can have more public views than impressions because of the estimated people watching together. Those figures therefore describe different things, rather than interchangeable measures of paid delivery.
Creator reports offer a separate comparison
YouTube’s content-performance documentation identifies a co-view rate that can be used to calculate a baseline without the additional shared-screen viewers. It also lists a co-viewed count in Insights Advanced Mode, available until May 2027, and defines unique reach as an estimate of individual people who saw a video or channel. Those measures answer different management questions.
For a brand reviewing performance, the distinction is between viewing activity, an estimated audience and a monetization measure. A public total does not supply all three. Comparisons need clear metric labels, reporting periods and identification of shared-screen estimates where the tools provide them, rather than treating every reported view as the same unit.
