Skip to main content

Pre-Match Alerts - Similar Users

Blair Matheson
  • Updated

The Pre-Match Alerts Similar Users feature leverages the power of collaborative behavior by identifying users with similar betting patterns. By recommending betting opportunities based on what similar users are betting on, you can create highly targeted campaigns that resonate with your customers and drive engagement.

Functionality

This feature uses Jaccard similarity to analyze users' betting histories, identifying customers with similar behaviors. It then recommends events based on those similar users’ preferences while also allowing you to exclude certain sports categories.

  • Behavior-Based Recommendations: Uses the intersection over union (Jaccard similarity) of user histories to find similar users.
  • Targeted Alerts: Suggests events based on what similar users are betting on.
  • Exclusion Options: Supports excluding specified sport categories from recommendations.

How it Works

  1. Identifies customers with similar betting patterns.
  2. Spots which events similar users have recently bet on.
  3. Matches recommendations with customers' favorite betting markets.
  4. Ranks event options based on popularity and odds.

Visualizing the Process

Flowchart (2).png

Configurable Parameters & Terminology

The following parameters are configurable to help tailor the recommendations:

  • Active Users Window: Days to analyze active users.
  • History Window: Period for analyzing betting history.
  • Min Similarity: Set the minimum Jaccard similarity score (0–1) required between users for recommendation matching.
  • Min Start Time / Max Start Time: Time window for upcoming events.
  • Exclude Categories: Sport categories to exclude from recommendations.
  • Similarity strength: How closely a user’s behavior matches others.

Outputs & Data Points

When a recommendation is generated, the system produces the following outputs:

  • category_name: The event's category.
  • event_date_time: Scheduled time when the event occurs.
  • event_key: Unique identifier for the event.
  • userid: Identifier of the recipient.
  • event_name: The title of the event.
  • market_name: Name of the betting market.
  • selection_name: The recommended betting option.
  • decprice: Odds for the recommended selection.

Data Integration Requirements

For proper functionality, the feature integrates with the following data sources:

  • Bet Place (Mandatory): Key fields include user ID, event name, event key, category name, market name, selection name, selection key, and event date/time.
  • Inventory Snapshot: Key fields include eventKey, eventName, categoryName, marketName, selectionName, selectionKey, decPrice, and eventDateTime.

Related articles