How Contextual Continuity embeds work
Contextual Continuity embeds carry a viewer's context across pages, sessions, and surfaces, so each feed refresh builds on what came before instead of starting over.
A standard contextual embed responds to a single moment: the page, a search, or a location. A Contextual Continuity embed adds journey awareness, so relevance doesn't reset abruptly as the viewer moves through your site or app.
Journeys aren't linear. Continuity builds familiarity, reduces friction, and increases return engagement and session depth.
Contextual Continuity is part of Onsite. For the product as a whole, see Onsite: what it does and where to start. For the capability's settings and reports, see Contextual Continuity.
Where Contextual Continuity applies
Continuity works in three kinds of implementation:
- Standard community embeds. Learn more about embedding your social experience on your website.
- Contextual embeds configured through the SDKs. Learn more in Deliver contextual feeds with page, user, and location context.
- Placements whose feeds adapt across pages or sessions. Learn more in Build a placement with styles, sections, and variables.
The configuration steps live in those placement and embed guides. This page explains the behavior they share.
What continuity does
Contextual Continuity is Genuin's framework for keeping and evolving context signals across a viewer's journey. Genuin doesn't treat each embed load as an isolated event. A Contextual Continuity embed:
- Remembers earlier context signals: page topics, interactions, and interests.
- Adapts recommendations step by step instead of resetting the feed.
- Balances freshness with familiarity.
- Keeps the feed relevant across navigation, scroll depth, and sessions.
The feed reads as intentionally curated, not random.
How continuity works
Contextual Continuity embeds build on standard contextual feeds. For how those choose videos, see How contextual feeds choose videos.
On top of the contextual feed, a continuity layer governs how context is kept, decayed, and updated over time. Genuin evaluates three inputs:
- The current page or interaction context.
- Recent historical context.
- Engagement signals from the same session or earlier sessions.
Genuin merges these inputs into a rolling context state. That state informs every later feed refresh.
Context signals
Page and content history
Genuin tracks recently visited pages, categories, and content themes to understand where the viewer is heading. For example, a viewer moves from "Electric Cars" to "SUV Comparisons", or reads several articles in one editorial vertical.
Tracking this history prevents sudden topic shifts in the embedded feed.
Engagement memory
The embed accounts for how the viewer interacts with content:
- Videos watched or skipped.
- Reactions, comments, and reposts.
- Dwell time and repeat views.
Themes with high engagement get reinforced, and ignored themes are gradually deprioritized.
Session context
Within one session, the embed adapts as the viewer's intent becomes clearer. For example, early feed items are broader, and later items get more specific based on interaction patterns.
The feed learns within the session without identifying the viewer.
Cross-session continuity
When user identifiers are present, Genuin resumes its understanding of the viewer's context across visits. It keeps themes consistent while it introduces fresh content.
Governance controls keep this behavior configurable and privacy-safe. For how a signed-in viewer is recognized across your properties, see Unified identity: auto login, SSO, and the ID graph.
How ranking uses continuity
Contextual Continuity embeds rank videos by blending three signals:
- Relevance to the current context.
- Affinity with the viewer's historical context.
- Recency and popularity.
Genuin doesn't replace the context on every load. It applies weighted decay instead: older signals fade unless new behavior reinforces them. Feeds stay relevant without repeating themselves.
Maya, Genuin's orchestrator, ranks every feed. For the signals it ranks on, see Personalization and relevance: how Maya ranks every feed.
Where continuity works best
Contextual Continuity embeds work best on:
- Editorial or content-heavy sites.
- Commerce discovery and comparison journeys.
- Multi-step consideration funnels.
- Experiences for logged-out or partially identified viewers.
They're commonly placed on:
- Article pages.
- Category and collection pages.
- Product detail pages (PDPs) and post-purchase surfaces.
- Scroll-driven discovery sections.
Continuity in community embeds
Within the contextual framework of community embeds, continuity makes sure that:
- Communities and groups feel cohesive.
- Content discovery follows the viewer's intent over time.
- Feeds evolve logically as the viewer goes deeper into a topic.
Governance and brand safety
Contextual Continuity runs under the same controls as every Genuin embed:
- Brand-level and community-level scoping.
- Content restrictions and moderation rules.
- Configurable persistence and decay.
- Full auditability of recommendations.
People stay in charge: AI assists discovery, and you keep control of what your embeds show.
