About the Cold Start feature

This is for:


Coveo Personalization-as-you-go (PAYG) models rely on product vectors and embeddings, as well as on user session vectors to deliver personalized recommendations and search results for each customer. These models are built by identifying relationships between indexed products based on behavioural data.

For products receiving high traffic, PAYG models can generate meaningful representations of user behavior in the form of vectors, which are then embedded directly into the product catalog. However, when dealing with long tail (less popular) or new products, there may not be a sufficiently large sample size to establish precise vector representations.

To optimize your catalog’s coverage and enhance the accuracy of recommendations and search results, Coveo PAYG models incorporate the Cold Start feature which allows them to effectively handle both long tail and new products.

How does the Cold Start feature work?

Representative example of the Cold Start feature | Coveo

Coveo PAYG models effectively position products with high traffic volumes within the vector space by analyzing behavioral data. However, when a product has limited or no traffic data associated with it, PAYG models rely on the Cold Start feature to place it within the pre-existing vector space.

To achieve this, instead of depending solely on customer interactions with a product, the Cold Start feature uses the product’s attributes to generate a product vector and position it within the vector space.

For example, the Cold Start feature can use the product’s category, description, and brand to generate a vector representation of the product and position it near other products that share similar attributes within the vector space.


The Cold Start feature is automatically activated for Intent-Aware Product Ranking and Predictive Query Suggestions models.