How Commerce Max Selects an Ad?

Commerce Max uses a combination of data signals to decide which Sponsored Product ad to show for each available impression. The goal is to maximize both advertiser performance and retailer yield. This decision-making process relies on three key inputs: contextual information, campaign performance, and user behaviour.  

These inputs are processed through Criteo’s advanced prediction models to determine ad relevance and assign a Quality Score - a measure of how likely a user is to engage with a specific product ad.


How it works? 

  1. Commerce Max evaluates contextual information 

    • This includes the page context where the ad opportunity occurs- such as product category or search terms. 

  1. It reviews campaign performance data 

    • It looks at how your product has previously performed: 

      • Engagement history (clicks, impressions, conversions). 

      • The types of users and categories where it performed best. 

      • Specific line item performance. 

    • Products with stronger historical performance tend to be favoured. 

  1. It considers user-level behaviour (for offsite and cross-context) 

    • Commerce Max uses first-party retailer data to understand what the user searched, browsed, clicked on, or purchased. 

    • This helps ensure ad placements align with a shopper’s demonstrated intent.

  1. It calculates a Quality Score 

    • The Quality Score (predicted Click-Through Rate or pCTR) is the outcome of all the above signals. 

    • A higher score means a higher likelihood that a shopper will click on the ad. 

    • You cannot directly influence this score; it is derived from: 

      • Brand familiarity 

      • Historical product performance 

      • Shopper engagement data 

  1. Commerce Max runs the auction 

    • Among eligible ads, it selects the one with the best combination of bid and Quality Score. 

    • Even with a high bid, a product with low predicted engagement may not win the auction. 

  1. Ad is delivered if it fits the context and performs well 

    • Only ads that align with the page context, have strong historical performance, and resonate with user behaviour are served. 

    • This ensures both brand visibility and user relevance, improving the overall experience for the shopper and value for you.