|
September 24, 2026 | SIGN UP
Jim Cooper For brands trying to bubble up their videos via YouTube LLM citations, longer seems to be better. One possible explanation, as we report this week, is that longer videos are more likely to include things like chapter markers, which gives LLMs granular signals about the content and where specific topics appear. Descriptive, keyword-aligned titles can then provide additional context about what the video covers. These are all things brand marketers understand well – content in context. Deloitte Digital’s head of content, creator and AI, Jenny Kelly, is seeing this shift in how brands approach video, with content shaped in ways that make it easier for AI systems to discover and understand. That can mean using more natural language, incorporating studies and FAQ formats, she told us. “They are literally setting the content stage for what the LLMs are looking for and how they prioritize what it is.”
ADVERTISEMENT
Top stories
PHOTO VIDEO In Graphic Detail: LLMs keep citing YouTube in search results Over the weekend, the Olympics racked up 17 billion impressions, 2 billion video views, and 49 million posts, according to X. Over the weekend, the Olympics racked up 17 billion impressions, 2 billion video views, and 49 million posts, according to X.
Measuring CTV on last-touch alone tells only part of the story. Across hundreds of advertisers, 30% of campaigns are significantly undervalued by last-touch models alone.
Other things to know
|