UpStart Commerce Search
Smart Ranking

How Smart Ranking Works in NoChannel Search

3min

NoChannel Search’s Smart Ranking system uses sophisticated machine learning algorithms to analyze user interactions and optimize search results. Let's break down the process:

  1. User Performs a Search Query: A customer types "blue jeans" into your apparel store's search bar.
  2. Initial Results Retrieved: NoChannel Search fetches all products matching "blue jeans" based on basic relevance factors like text matching.
  3. Smart Ranking Model Analysis: The Smart Ranking model swings into action, considering multiple factors:
    • User click data: Which blue jeans have users clicked on most often?
    • Product popularity: Are certain styles of blue jeans trending?
    • Textual relevance: How closely do product descriptions match "blue jeans"?
    • User behavior: What have similar users clicked on when searching for jeans?
  4. Results Re-Ordering: Based on this analysis, the model assigns a relevance score to each product and re-orders the results.
  5. Display of Re-Ranked Results: The customer sees the re-ranked list, with the most relevant blue jeans appearing at the top.

Data Capture Flow for Smart Ranking

The Smart Ranking service relies on a comprehensive Data Capture Flow to understand user behavior and improve search results. This flow consists of several key events:

  1. Session Initialization
  2. Navigation Tracking
  3. Search Event Logging
  4. User Interaction Recording
  5. Cart Event Tracking
  6. Promotion Application Logging
  7. Order Placement Capture

Each of these events is timestamped and associated with the user's session, providing a complete picture of the user's journey. This rich dataset feeds into the Smart Ranking system, allowing for:

  • Detailed analysis of user search patterns
  • Understanding of which products are most relevant for specific queries
  • Insights into how users navigate from search to purchase

By capturing this comprehensive set of user interactions, our Smart Ranking service can continuously learn and improve, offering increasingly relevant search results and enhancing the overall shopping experience.

For more details, go to our Data Capture document.