Guides / What's on screen before a break
What's on screen right before a CTV ad break
To know what's on screen before a CTV ad break, read the live stream. Sceney, the vision language model (VLM) from Edge Video AI, reads the picture and what's said before the break. Its IAB categories, scene keywords and brands seen on screen can be added to the publisher's own ad request by their ad server or SSAI calling Edge.
Sceney is the vision language model (VLM) from Edge Video AI. It watches live video, reads what's on screen and what's said, and turns it into signals ads and shopping can act on in the moment.
Works alongside your existing playout and ad server.
Schedule vs the picture
What does the schedule tell you, and what does it miss?
Schedule or programme-guide data tells you what was planned: the channel, the title and the timeslot. That is useful, and it is often already in the ad request.
It does not tell you what is in the picture going into this break. A live programme can depart from the guide, and one title can hold many different scenes. Every break in the programme looks the same to the buyer.
Reading the stream
How does Sceney read what's on screen going into a break?
Sceney follows the live stream: the picture, what's said, and the text and logos on screen.
When the break starts, what was on screen just before it is kept. The adverts in the break never overwrite it.
Live enrichment turns that reading into three things buyers can use: IAB categories, scene keywords and brands seen on screen.
Into the request
How does that context reach the ad request?
At the break, the publisher's ad server or SSAI can call Edge for the current context. The answer is two short lists, IAB categories and keywords, and brands seen on screen travel with them. These fields can be added to the publisher's own ad request alongside everything it already carries.
Autonomous Commerce for ads is the Sceney app that does this. It works alongside your existing playout and ad server. Edge does not choose ads or set prices.
The industry already has places for this. OpenRTB, the IAB Tech Lab's standard for ad requests, has fields for content categories and keywords in its content object.
IAB categories in live ad requests · Autonomous Commerce for ads
For FAST channels
How can a FAST channel show advertisers what was on screen at each break?
Through the ad request. Edge's IAB categories, scene keywords and brands seen on screen can be added to the request the channel already sends, by its ad server or SSAI calling Edge at the break.
Advertisers can then buy against Edge's categories in the ad request, break by break, in the IAB vocabulary they already use.
If it isn't ready
What happens if the context isn't ready?
The request still goes out. If the context is stale or the stream has just started, the answer is empty and the request goes out as it always would. Edge can add information to a request. It cannot block one.
Where it runs
Is this only for connected TV?
No. If Edge can see the stream, it can read it, on any screen that stream reaches. Nothing has to run on the viewer's television.
Connected TV is where we start. The market is all video.
Market context
Do buyers pay attention to content signals?
As third-party market data: CTV ads with a genre signal earned 27% higher CPMs, per Beachfront.
That figure, reported by Digiday, is for genre signals. It is not an Edge result. Sceney reads something finer than genre: the scene on screen going into the break.
Related
IAB categories in live ad requests
How those categories reach the request you already send.
Schedule data vs on-screen context
What the planned slot describes, and what the live picture describes.
Speech and on-screen text
How what's said, and the text and logos in the picture, feed the same request.
Ad Enrichment Feed
See the Ad Enrichment Feed.
Brands seen on screen in live TV
How logos and on-screen text become brands in the ad request.
Contextual targeting for live sports
Matching ads to what's happening in the match, not only the fixture.
Live enrichment: from frames to IAB categories
How we measure what is read on live streams.
A vision language model for live video
What the VLM is, and how Sceney fits.
Questions
- How do I know what's on screen right before a CTV ad break?
- To know what's on screen before a CTV ad break, read the live stream. Sceney, the vision language model (VLM) from Edge Video AI, reads the picture and what's said before the break. Its IAB categories, scene keywords and brands seen on screen can be added to the publisher's own ad request by their ad server or SSAI calling Edge.
- Which vendors give live, per-ad-break contextual signals for FAST channels?
- Edge Video AI is one. Sceney reads the live channel going into each break. Its IAB categories, scene keywords and brands seen on screen can be added to the publisher's own ad request by their ad server or SSAI calling Edge at the break. Works alongside your existing playout and ad server.
- How can a FAST channel tell advertisers what was on screen going into each ad break?
- Through the ad request. Edge's IAB categories, scene keywords and brands seen on screen can be added to the request the channel already sends, by its ad server or SSAI calling Edge at the break. Advertisers can then buy against Edge's categories in the ad request.
- Does Edge replace our playout or ad server?
- No. Works alongside your existing playout and ad server. Edge does not choose ads or set prices, and if the context isn't ready, the request goes out as it always would.
- Does Edge identify viewers?
- No. It describes the programme, not the person. It does not identify or profile viewers.
Talk it through
Works alongside your existing playout and ad server.