Skip to main content

Least Active Day

Blair Matheson
  • Updated

The Least Active Day feature recommends one of multiple personalization types of game recommendations on each user’s least active day of the week.

Functionality

  • Identifies for each user the least active day.
  • Recommends a game to the user.
  • Uses a data-driven approach to ensure relevant and engaging recommendations.

How it Works

  1. Identify Audience:
    • Analyze User Activity: Examine historical data to determine each user’s activity levels on different days of the week in the past n_weeks weeks.
    • Identify Least Active Day: Calculate the proportion of a user’s activity on each day relative to their total activity and compare to the overall user‑base distribution; choose the day where a user’s activity is lowest (in a tie, pick the more popular day).
    • Select Today’s Least Active Users: Filter users whose least active day corresponds to the current day of the week.
  2. Determine Favorite Games: Use the favorite game model to analyze user gameplay history and engagement scores, extract the highest‑ranked game for each user—and include game codes, names, and images for personalized messaging.
  3. Determine Recommended Games: Analyze user gameplay patterns and engagement scores; use predictive models to recommend a game that aligns with the user’s preferences.
  4. Determine Favorite Games: Analyze user gameplay history and engagement scores again to confirm the single favorite game recommendation per user.
  5. Determine Popular Games: Identify the most played games from recent trends, filter by category if applicable, and select the top game by unique players.
  6. Personalized Recommendations: Deliver one game recommendation per user based on the chosen recommendation type, including game codes, names, and images.

Configurable Parameters & Terminology

  • Weeks to Analyze: Number of weeks to calculate each user’s least active day of the week.
  • Game Category: Filter for game categories to recommend.
  • Recommendation Model: Options for game recommendation types.

Outputs & Data Points

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

  • userid: Identifier of the recipient
  • game_code: Game identifier
  • game_name: Name of favorite game
  • imageurl: Favorite game image URL (JSON format for optiXtra)

Data Integration Requirements

For proper operation, this feature integrates with the following data sources:

  • GAME_PLAY events (Mandatory): Key fields: userid, game_code, game_name.
  • Inventory (Mandatory): Key fields: game_name, gamecode, imageurl.
  • Reference File (Mandatory): Key fields: game_name, gamecode, game_category, imageurl.

Related articles