Multi-Strategy Algorithms

Prev Next

Multi-Strategy algorithms combine multiple recommendation algorithms to improve relevance and diversity. Leveraging the strengths of each algorithm, they deliver more balanced and effective recommendations. 

The algorithms covered under Multi-Strategy Algorithms are:

Mixed Strategy

The Mixed Strategy algorithm employs various algorithms for different slots within the Recommendation Widget. This approach enables the simultaneous use of multiple algorithms within a single recommendation widget on your website.

  • Available Channels: Depends on the algorithms used
  • Page Type: All Pages, Product/Article Pages, Category Pages, Cart Page
  • Example Use Case: Promoting products provided from a different range of Smart Recommender algorithms
  • Fallback Algorithms: The fallback algorithm of each slot's algorithm is applied.
  • Prerequisites: Depends on the selected algorithms
  • Maximum Number of Products to Recommend (in the same variant): 15 

Chef

The Chef algorithm, an automated recommendation algorithm, dynamically determines the optimal mixed strategy combination by assessing algorithms such as the most popular items, top sellers, user-based, highest discounted, new arrivals, and trending items. This iterative testing process aims to enhance conversion rates and deliver the most effective recommendations.

  • Available Channels: Web Smart Recommender
  • Page Type: Home page
  • Example Use Case: Considering the most popular items, top sellers, user-based, highest-discounted, new arrivals, and trending items algorithms, the Chef evaluates each algorithm's recommendations one by one by calculating the precision of each slot and eventually recommends products with the highest precision. 
  • Fallback Algorithms: The fallback algorithm of each slot's algorithm is applied.
  • Prerequisites: 
    • It evaluates the most popular items, top sellers, user-based, highest discounted, new arrivals, and trending items algorithms, whichever is enabled for your account. If only 3 of those are activated, Chef creates its combination based on those three open algorithms.
    • The Chef works with 60 days of data. For better results, we highly recommend keeping Smart Recommender open for at least 2 months. 
    • The Chef also works for new accounts (with data less than 60 days) by providing a default strategy derived from best practices. However, its results will improve once more data is collected.
  • Maximum Number of Products to Recommend (in the same variant): 15
The Shuffle recommended products feature is not available for the Chef algorithm.