TL;DR
- YouTube introduced views co-viewed in YouTube Analytics on 5 September 2026 through a Creator Insider video hosted by creator liaison Rene Ritchie.
- The estimate is built from three named model inputs, which are demographic patterns, video genre and viewing times. No error band, retraining cadence or third-party audit was disclosed.
- Payout mechanics are untouched. YouTube said revenue sharing, AdSense payouts and Partner Program eligibility still run on engaged views and qualified views.
- Worked example from the video: a household of three watching one video twice produces six co-viewed views and three unique reach viewers, unique reach having arrived in May 2026.
- Public view counts on the watch page are unchanged, and the video did not say whether advertisers will see the metric in Google Ads reporting.
YouTube announced a YouTube Analytics metric called views co-viewed on 5 September 2026. The metric estimates the total number of people watching a video on a television screen, replacing an arrangement in which every TV device playback was viewed as a single view, so a household of three people watching together can now register as three estimated views.
The figure is produced by statistical modelling rather than by counting people. Public view counts on the watch page do not change, and creator payouts do not change. What changes is the sponsorship conversation, because a creator can now open YouTube Analytics and show a television audience larger than the same account reported a week earlier.
A quoted YouTube reach figure can now come from at least four different counters
With views co-viewed added on 5 September 2026, one YouTube video has several separate audience counters attached to it: the public view count on the watch page, engaged views, qualified views, unique reach, and now views co-viewed. Each answers a different question, and only some of them decide money.
That matters to anyone reading a creator media kit, because a slide saying a channel reaches a given number of people no longer identifies which counter produced that number. The distance between the smallest and the largest of them, for the same video in the same week, can be wide. Using the worked example YouTube gave in the Creator Insider video, a household of three watching one video twice produces six co-viewed views and three unique reach viewers.
The table below sets out the counters named in the announcement, what each one measures, and whether YouTube said it feeds creator payouts.
| Counter | What it counts | Does it drive payouts? |
|---|---|---|
| Public view count | The figure shown on the watch page. YouTube said it is unchanged by this release. | Not named as a payout basis |
| Engaged views | One of the two metrics YouTube said revenue sharing, AdSense payouts and Partner Program eligibility continue to run on. | Yes |
| Qualified views | The second metric YouTube said payouts and Partner Program eligibility continue to run on. | Yes |
| Unique reach | Arrived in Advanced Analytics in May 2026. Three people watching the same video twice count as three unique reach viewers. | Not named as a payout basis |
| Views co-viewed | Announced 5 September 2026. A modelled estimate of the total people watching on TVs. Three people watching the same video twice produce six co-viewed views. | No, payouts run on engaged views and qualified views |
Before 5 September 2026, one television counted as one viewer
YouTube described the arrangement it is replacing in plain terms: every TV device playback was viewed as a single view. A television set is one device running one session, so three people on a sofa produced one view, and YouTube Analytics had no way to represent the other two.
The mechanism underneath the change is a substitution. Where the old count measured a device, the new estimate infers people from patterns YouTube can already see. Nothing new is being observed in the living room. The same signal is read through a model that outputs a larger number, and the size of that increase rests on assumptions the video did not publish.
YouTube named three model inputs on 5 September 2026 and disclosed nothing else about the model
The Creator Insider video said the estimate comes from statistical modelling based on demographic patterns, video genre and viewing times, used to predict when people are watching together. Those are the three inputs YouTube named. The write-up of that video by PPC Land is the readable record of what was said, since the announcement itself was a video hosted by creator liaison Rene Ritchie on a channel that lists 910,000 subscribers.
Nothing else about the model was disclosed. The video did not describe the model itself, did not publish error bands or an accuracy figure, and did not say how often the model is retrained. Analysis: a prediction with three named inputs and no published error term is still a prediction, and the honest way to read a co-viewed figure is as a modelled range that YouTube has chosen to display as a single number.
Views co-viewed and unique reach answer different questions
YouTube distinguished views co-viewed from unique reach, which arrived in Advanced Analytics in May 2026, and gave a worked example in the 5 September 2026 video: a household of three watching the same video twice produces six co-viewed views and three unique reach viewers.
Read together, the two numbers bracket the same audience. Unique reach counts a person once, so it answers how many humans saw the video. Views co-viewed counts each person on each play, so it answers how many human viewings the video generated on television screens. A campaign brief asking for reach wants the first number. A creator arguing for a higher fee has an incentive to quote the second.
Public view counts and creator payouts are unchanged
YouTube said public view counts on the watch page are not affected by views co-viewed, and that revenue sharing, AdSense payouts and Partner Program eligibility continue to run on engaged views and qualified views. The new metric is a reporting change inside YouTube Analytics, not a change to what YouTube pays for.
That boundary is the most useful single fact in the announcement. It says YouTube is willing to model an audience for the purpose of describing it, while keeping the money on metrics it counts directly. A buyer can apply the same standard: use the modelled number to understand the shape of television viewing, and use directly counted metrics when something is being bought or paid for.
YouTube framed the metric for brand deals, which puts a modelled multiplier on one side of the table
The Creator Insider video presented views co-viewed as something creators can use when negotiating brand deals. YouTube also repeated its claim that viewers watch an average of over one billion hours of YouTube on television screens every day.
Analysis, not a claim from the source: that combination places a platform-supplied multiplier inside the creator's own dashboard, with no external audit attached, at the moment the creator is quoting a price. None of that is improper. It does mean a buyer has to ask a question that did not exist before 5 September 2026, which is which counter produced the number on the slide.
Who this affects, and who it does not
Analysis: the metric matters most to channels with a large share of watch time on television screens, and least to channels watched almost entirely on phones, where co-viewing is rare and the estimate should sit close to the old count. Brand teams evaluating creator proposals are affected immediately, because the numbers in front of them can move with no change in audience behaviour. Advertisers buying YouTube inventory are not affected in any way YouTube described, since the announcement said nothing about advertiser-facing reporting. Payout processes are not affected at all, on YouTube's own statement.
What the Creator Insider video did not say
The 5 September 2026 announcement left several things open, and they are worth stating plainly rather than filling in:
- No description of the model behind the estimate.
- No accuracy figure and no error band.
- No retraining cadence for the model.
- No third-party audit of the estimate.
- No country-level availability detail, so whether every channel everywhere sees the metric was not addressed.
- No statement on whether advertisers, as opposed to creators, will see views co-viewed in Google Ads reporting.
Anything past that list is inference. Where this article reasons beyond the source, it says so.
What this means for marketers in Thailand
The source did not discuss Thailand, so this section is reasoning rather than reporting. Thai households are often large and multi-generational, and watching YouTube on the main television set is ordinary family behaviour, so a co-viewing multiplier should read higher here than in markets dominated by single-occupancy homes. A Thai creator's co-viewed figure may therefore climb further above their old television view count than a European creator's would.
Influencer pricing in Thailand is not well benchmarked to begin with. A platform-supplied number that raises television audience estimates without an audit trail hands creators a fresh argument for a higher fee, and gives brand teams no independent way to check it. The workable response is not to distrust the metric but to write the metric name into the paperwork, so a reach commitment in a contract refers to a defined counter rather than to whichever screenshot was pasted into a deck. Teams already running YouTube advertising alongside creator deals have an advantage here, because paid reporting gives them a second reference point for how the same audience behaves.
What to check in your own reporting this week
- Ask every creator proposal which metric produced the reach figure, and get the answer in writing.
- Compare a channel's unique reach with its views co-viewed for the same period before agreeing a rate, because the two numbers bracket the audience.
- Check whether any internal dashboard or client report pulls a YouTube figure that has now changed definition, and label the metric on the chart rather than in a footnote.
- Keep contracted deliverables on directly counted metrics where you can, since YouTube itself keeps payouts on engaged views and qualified views.
- Do not restate historical television performance using the new metric without saying so, because a year-on-year comparison across a definition change is not a comparison.
FAQ about the views co-viewed metric
Does views co-viewed change my video's public view count?
No. YouTube said on 5 September 2026 that public view counts on the watch page are unchanged, and that views co-viewed sits in YouTube Analytics.
Does views co-viewed affect AdSense payouts or Partner Program eligibility?
No. YouTube said revenue sharing, AdSense payouts and Partner Program eligibility continue to run on engaged views and qualified views, which are counted rather than modelled.
How accurate is the co-viewing estimate?
The source did not state it. The Creator Insider video named three model inputs, which are demographic patterns, video genre and viewing times, but published no accuracy figure, no error band and no third-party audit.
Is views co-viewed available in Thailand?
The source did not state country-level availability. The announcement described the metric and its inputs without saying which markets or channel types see it first, so a Thai channel should check its own Analytics rather than assume.
Will advertisers see views co-viewed in Google Ads reporting?
The source did not say. The announcement went to creators through Creator Insider and described the metric inside YouTube Analytics, with no statement about advertiser-facing reporting.
Where to go from here
If your team is about to sign creator deals for the fourth quarter, the cheapest fix available this week is a definitions line in the contract template, naming the counter behind every reach commitment. If you would rather have someone read the media kits with you and set one measurement standard for every deal, that is work we do alongside paid social campaigns. Tell us the numbers you have been quoted and we will start there.
