Contextual Embeds: How They Work and How to Configure Them

Contextual Embeds allow you to display dynamic, personalized video feeds directly on your website or app, based on the user's real-time context. This document walks you through what contextual feeds are, how they work, and how to configure them using Genuin’s SDKs.

What Are Contextual Feeds?

A contextual feed is a personalized video stream generated using signals from the user’s current environment, such as:

  • Page context (content or category of the page)
  • Search keywords
  • User attributes (age, profile data, interests)
  • Geolocation (latitude & longitude)

By processing these signals, Genuin recommends videos that are more relevant to the user at that specific moment. The result is a real-time, adaptive, and highly engaging video feed.

How Contextual Feeds Work

Genuin’s contextual recommendation engine evaluates multiple context types - each represented as its own structured field. At any given time, the feed uses the highest-quality available context to generate video recommendations.

Supported Context Inputs

  • Webpage Context
    The page the user is viewing (e.g., “SUV listings”, “Car Loans Guide”).
  • Search Query Context
    Keywords entered by the user.
  • User Profile Context
    Age, bio, stated preferences, interests.
  • Geographic Context
    Latitude and longitude, often derived from IP or device.

Video Selection and Ranking Logic

Video recommendations are generated using a multi-stage process:

1. User Profile–Based Feed Generation

  • Retrieves videos that match the user’s profile signals.
  • If no matches exist (cold-start), the system falls back to contextual or geo-based selection.
  • Engagement-based ranking:
    • Shares (highest weight)
    • Sparks (medium)
    • Views (lowest)

2. Context-Based Feed Generation

  • Processes text-based context (page context, searches, bio) to extract keywords, topics, and user intent.
  • Incorporates geolocation for localized recommendations.
  • Executes an OpenSearch query using these extracted signals.

3. Feed Ranking & Final Output

Video ranking is computed using a weighted score:

index.html
Final_score = 0.34 * Context_score 
            + 0.33 * Recency_score
            + 0.33 * Popularity_score

Videos are then presented in descending order of final score.

This ensures each user receives highly relevant, recent, and popular content.

Contextual Feed Parameters

To enable contextual recommendations, pass the following payload:

index.html
{
  "brand_id": int,
  "communities": [
    {
      "community_uuid": "string",
      "groups": ["string", "string", "string"]
    },
    {
      "community_uuid": "string",
      "groups": ["string", "string", "string"]
    }
  ],
  "user_uuid": "string",
  "user_interest": ["string", "string", "string"],
  "page_session": "string",
  "genuin_user": true,
  "limit": int,
  "contextual_feed": true,
  "user_weight": float,
  "recency_weight": float,
  "popularity_weight": float,
  "page_context": {
    "page_context": "string",
    "compare_with": "video/community/brand/group/location",
    "location": [lat, long],
    "location_radius": int
  }
}

How compare_with Works

compare_withBehavior
communityMatches videos from communities with similar descriptions
brandMatches videos linked to similar brands
videoFinds contextually similar videos
locationFilters videos within a specified radius

If contextual_feed = false, the system falls back to standard (non-contextual) video selection logic.

Contextual Embed Activation Flow

  1. Contextual feed flag is checked.
  2. Videos filtered by brand_id.
  3. A context vector is generated from the provided inputs.
  4. Videos receive a contextual matching score.
  5. Videos are ranked and returned in descending relevance.

This enables a hyper-personalized video experience on your website or app.

How to Configure Contextual Feeds (Web, iOS, Android)

You can integrate contextual feeds using Genuin’s SDKs. Follow the platform-specific documentation here:

These guides provide step-by-step instructions, code examples, and configuration patterns tailored to each platform.

Summary

Contextual Embeds help you deliver real-time personalized video recommendations based on user intent, behavior, and location. With simple configuration options and powerful multi-context ranking, you can instantly enhance on-site engagement and user experience across all your digital surfaces.

COMPANY

Genuin Footer