Behaviour and similarity signals
The intelligence behind every recommendation
Overview
Liva 7’s personalisation engine relies on two powerful signal types - behavioural and similarity - to understand what shoppers are doing and how products relate to one another. These signals are the foundation of real-time, relevant recommendations across your storefront.
This tab gives you visibility into how these signals are performing and whether they’re actively contributing to your personalisation strategy.
Behaviour Signals
Real-time insights from shopper activity
Behaviour signals track how visitors interact with your store. These include:
- Browsing behaviour
Tracks page views, product clicks, and time on site - Cart behaviour
Captures add-to-cart and remove-from-cart events - Purchase history
Uses completed orders to inform future recommendations
Each signal displays a status:
- Active: Data is flowing and being used in real time
- Waiting for data: Signal is enabled but hasn’t received enough activity yet
- Inactive: Signal is not currently being tracked
Browser behaviour helps Liva 7 understand what shoppers are interested in, even before they add anything to cart.
Similarity Signals
Understanding product relationships
Similarity signals help Liva 7 recommend products that are visually, functionally, or behaviourally related. These include:
- Similarity patterns
Based on co-viewed, co-purchased, and co-carted product behaviour - AI vector analysis
Uses machine learning to understand product attributes, descriptions, and visual features
These signals are especially useful when:
- Behavioural data is limited (e.g. new visitors or products)
- You want to surface alternatives or complementary items
- You’re building “Similar Items” or “Complete the Look” experiences
Similarity signals identify patterns in how products are browsed or bought together. AI vector analysis goes beyond tags and categories to find deep product relationships using embeddings.
True intelligence
Behaviour and similarity signals are what make Liva 7 truly intelligent. By combining real-time shopper activity with deep product relationships, the engine can deliver recommendations that feel intuitive, timely, and tailored - even for first-time visitors.
