User Profile

A user profile contains behavioral information about a given user, such as their actions history, preferred brands and product categories. This information is leveraged internally by Coveo services, such as the Coveo for Salesforce User Actions Component or Coveo Machine Learning (Coveo ML) models to provide more personalized and task-oriented recommendations.

To do so, Coveo uses all users’ actions recorded by Coveo Usage Analytics (Coveo UA) and applies statistical and machine learning techniques to analyze behavioral data to better understand user interests, preferences, and intents.

About the User Profile

A user profile is divided into multiple user profile dimensions, each one of them representing a user’s activity, characteristic, or interest (e.g. last items viewed, preferred brands, topics of interest).

These dimensions can either be calculated and updated in real time, or according to a specific usage analytics data exportation period.

Important

User profile dimensions must be enabled to take advantage of a user profile in Coveo services. Contact us to enable these dimensions.

Current Capabilities

Currently, the dimensions that are required to build a user profile must be manually activated to be effective in Coveo services. You must Contact us to enable these dimensions.

User profiles are used by the following services:

Coveo ML Query Suggestions

For all Coveo ML Query Suggestions (QS) and Predictive Query Suggestions (PQS) models, the user topics dimension is automatically activated, which means that a given QS or PQS model adapts its recommendations according to the user’s main topic of interest.

Coveo ML Product Recommendations

Currently, Coveo ML Product Recommendations (PR) models use the favorite brands, favorite categories and history of actions dimensions to constantly adapt the recommendations according to the user’s navigation history and product preferences.

Coveo ML Content Recommendations

Coveo ML Content Recommendations (CR) models can leverage the history of actions dimension to personalize the recommendations.

By default, Coveo ML CR models use the information of the actionsHistory query parameter, which is fed by a browser cookie that stores the user’s session navigation history.

When leveraging the history of actions dimension, Coveo ML CR models don’t rely on the user’s actionsHistory to obtain their session’s navigation history, removing the need for users to both store this information in a browser cookie and send it to the model.

Coveo ML CR only uses the user’s visitor ID to query the user profile database which stores their session navigation history.

User Actions Component

The Coveo for Salesforce User Actions Component uses the history of actions dimension to record the history of actions performed by a given user.

This allows support agents to see, in the Salesforce Lightning console, up to the last 2,000 actions performed by an end user across any Coveo page or component, which gives support agents crucial insights on the end user’s task.

User Profile Dimensions

User profiles contain information learned from the following dimensions.

Important

User profile dimensions must be enabled to take advantage of a user profile in Coveo services. Contact us to enable these dimensions.

History of Actions

A user profile is constantly being updated according to the actions a user performed.

This dimension records the last 2,000 actions performed by the user, including searches, clicks, and consulted items.

Coveo services such as the Coveo for Salesforce User Actions Lightning Component, Product Recommendations, and Content Recommendations ML models use this information to personalize the user experience in real time according to the actions performed by the user in their current session.

Example

During a shopping session on a Coveo-powered pet supplies commerce interface, an unauthenticated user performs the following actions:

  1. Performs the dog collar query.

  2. Selects the red value of the Color facet to refine the query.

  3. Clicks one of the collars that appears on the results list.

  4. Adds the collar to the shopping cart.

By comparing the actions performed in the above session with those of other users who performed similar actions, a Coveo ML PR model can deduce that the user is currently looking for dog products.

Therefore, the model can recommend products that have been purchased or viewed by other similar users who also purchased the same dog collar the current user added to their cart.

Favorite Brands

For a user browsing Coveo-powered commerce interfaces, the user profile is constantly being updated according to the user’s preferred brands.

Product Recommendations ML models use this information to recommend products of brands that the user is most likely to buy.

Example

During a shopping session on a Coveo-powered electronics retailer commerce interface, an unauthenticated user performs the following actions:

  1. Performs the 55-inch TV query.

  2. Selects the ACME value of the Brand facet to refine the query.

  3. Clicks one of the televisions that appears on the results list.

  4. On the Frequently viewed together product recommendation interface that appears on the product page, the user clicks a television of the EMCA brand.

With this information, a Coveo ML Product Recommendations model can deduce that the user is more likely to purchase ACME and EMCA products.

Therefore, the model adapts its recommendations to show products of the ACME and EMCA brands.

Favorite Categories

For a user browsing Coveo-powered commerce interfaces, the user profile is constantly being updated according to the user’s preferred categories.

Product Recommendations ML models use this information to recommend products of categories that the user is most likely to buy.

Example

During a shopping session on a Coveo-powered furniture commerce interface, an unauthenticated user performs the following actions:

  1. Performs the king-size mattress query.

  2. Clicks one of the mattresses that appears on the results list.

  3. On the Frequently bought together product recommendation interface that appears on the product page, the user clicks on the ACME King Platform Bed recommended product, and then adds the product to their shopping cart.

  4. Performs the dresser query.

With this information, a Coveo ML Product Recommendations model can deduce that the user is currently more likely to purchase bedroom furniture.

Therefore, the model adapts its recommendations to provide products of the bedroom category.

User Topics

Query Suggestions ML models adapt their recommendations according to the distribution of topic clusters learned for a given user.

Example

In a search interface in which a Coveo ML QS model is integrated, the query completion suggestions will differ depending on the user’s profile. This level of personalization provides your users with a more intuitive experience as the model’s suggestions are adapted to their preferences and search intents.

user profile in query suggestion model