Depending on your campaign objective, we apply one of three optimization algorithms to prioritize the best-performing SKU:

• Click Optimizer: Ideal for driving traffic to product pages
• Conversion Optimizer: Focuses on maximizing purchases
• Revenue Optimizer: Targets the highest return on ad spend

Each model adjusts how the quality score is calculated and how bids are optimized to align with your campaign’s success metric.

The quality score is a predictive value generated by the Criteo Retail Media engine. It estimates how likely a user is to engage with a product ad, whether by clicking, converting or generating revenue. This is composed of three key components:

• Predicted click-through rate (pCTR): The likelihood a user will click on the ad
• Predicted conversion rate (pCVR): The likelihood that a click will lead to a purchase
• Predicted sales amount (pSA): The expected revenue from that purchase

These elements are combined to form a quality score that reflects both user intent and product relevance.

With contextual relevance and intent signals, the algorithm also considers:

• Contextual relevance: Whether the product matches the shopper’s browsing behavior and the page context
• Intent signals: Behavioral cues that indicate a shopper’s likelihood to engage with a specific product

All eligible SKUs enter an auction where each is assigned a rank value calculated as Rank = Quality Score x CPC Bid. Below are the three different optimization models that can be selected based on your campaign's individual goals.

Click Optimizer

The goal is awareness. That achieves as many clicks as possible.

1. A shopper visits BestBuy.com to buy a camera.
2. They type “camera” in the search bar.
3. Two products with the same relevancy score enter the auction.
4. Camera A is bidding at $0.80, and Camera B is bidding at $0.75.
5. Camera A wins the placement because it bid higher.

Conversion Optimization

The goal is Conversions - to sell as many units as possible. The algorithm considers product bids and conversion rates to determine which product will win the placement.

1. A shopper visits BestBuy.com to buy a camera.
2. They type “camera” in the search bar.
3. Two products with the same relevancy score enter the auction.
4. Camera A is bidding at $0.80, and Camera B is bidding at $0.75.
5. Based on historical shopper data, the algorithm knows Camera A has a 5% predicted conversion rate and Camera B has a 6% conversion rate.
6. Camera A: $.80 x 2% CTR X 5% = .0008 Camera B: $.75 x 2% CTR X 6% = .0009
7. Camera B has a higher effective CPC than Camera A, so it wins the placement.

Revenue Optimizer

The goal is a return on ad spend (ROAS). This achieves the highest attributes of sales possible. The algorithm considers product bids, conversion rates and price points to determine which product will win the placement.

1. A shopper visits BestBuy.com to buy a camera.
2. They type “camera” in the search bar.
3. Two products with the same relevancy score enter the auction.
4. Camera A is bidding at $0.80, and Camera B is bidding at $0.75.
5. Based on historical shopper data, the algorithm knows Camera A has a 5% predicted conversion rate and Camera B has a 6% conversion rate. Camera A costs $60 and Camera B costs $40
6. Camera A: $0.80 x 2% CTR X 5% x $60 = .048 Camera B: $0.75 x 2% CTR X 6% x $40 = .038
7. Camera A has a higher effective CPC than Camera B, so it wins the placement.

Optimize campaign performance for stronger results

Combining predictive modeling with real-time auction logic ensures that the most relevant and competitive products are surfaced to shoppers. Whether your goal is to drive clicks, boost conversions or maximize revenue, our algorithmic approach adapts to meet your campaign objectives, delivering smarter placements and better outcomes.