---
title: "Predictive Segments: Attribute Affinity"
slug: "predictive-segments-attribute-affinity"
description: "Identify Attribute Affinity in Insider One Predictive Segments to target users based on product attributes like brand, category, or color preferences."
tags: ["attribute affinity", "predictive segmentation", "predictive segments"]
updated: 2026-04-12T18:27:49Z
published: 2026-04-12T18:27:49Z
canonical: "academy.insiderone.com/predictive-segments-attribute-affinity"
---

> ## Documentation Index
> Fetch the complete documentation index at: https://academy.insiderone.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Predictive Segments: Attribute Affinity

User-based targeting is becoming increasingly important in digital marketing, especially when identifying users based on their affinity for product-level attributes. This approach supports personalized recommendations and enables segmentation based on user engagement and likelihood of conversion.

The Attribute Affinity model collects historical data from the past 30 days based on user interactions with products, such as views, add-to-cart actions, and purchases. It determines each user's affinity for specific attributes (e.g., brand, category, or color) and segments users accordingly, allowing marketers to deliver highly targeted recommendations and campaigns.

![](https://cdn.document360.io/c6df4583-da94-4cb2-bb8a-be0cbdd11109/Images/Documentation/Screen%20Shot%202022-01-18%20at%2015.51.08.png)

## Model Details

The model evaluates user behavior by tracking product interactions and identifying preferences for specific [attributes](https://academy.insiderone.com/docs/product-catalog-attribute-list) such as brand, color, or size. This segmentation enables more effective personalized recommendations and targeted marketing.

- **Segment update frequency**: Daily
- **Data span**: Last 30 days
- **Calculation time**: 30 days
- **Model type**: Clustering algorithm or multi-class classification model

## Data Used

The following data points are analyzed to infer user preferences and build attribute-based segments:

- Product page views
- Purchases
- Specific interactions with different product types
- Filter preferences (e.g., color, size, brand)

## Key Features

The following features are used to evaluate and segment users based on their interactions with product attributes:

- Filter interactions
- Browsing history and category preferences
- Time spent viewing specific products

This model supports use cases like surfacing preferred colors at the top of catalogs or highlighting a user's favorite brands, helping increase engagement and conversions.

## Related

- [Predictive Segments](/audience-predictive-segments.md)
- [Use Predictive Segments](/use-predictive-segments.md)