Predictive User Engagement provides a simple architecture that automates the process of making predictive recommendations based on user activity in Amazon Personalize, and updating Amazon Pinpoint endpoints with those recommendations.

This Guidance is designed to provide a simple architecture to demonstrate how to use ML to make product recommendations and automatically update your endpoints and segments. You can build upon this architecture for a variety of use cases.

Overview

The diagram below presents the architecture you can build using the example code on GitHub.

Predictive User Engagement | Architecture Diagram
 Click to enlarge

Predictive User Engagement architecture

The code deploys an AWS Lambda function that ingests user activity data from an application. The function sends that data to Amazon Personalize, which runs a machine learning (ML) model on the data to identify patterns. Amazon Personalize generates a personalized ranking of recommended items for each user ID.

The Lambda function retrieves the personalized rankings and sends them to Amazon Pinpoint, which uses these recommendations to automatically update endpoints that belong to your segments based on how the personalized ranking matches your segment filters. For example, if a customer who you were sending messaging on product A now shows a preference for product B based on recent activity, this Guidance will automatically update the customer endpoint to move the endpoint from the segment that receives product A messaging to the segment that receives product B messaging.

You can also set campaigns to send personalized, timely, and relevant messages to the segments this Guidance updates. You can choose to send messages immediately, in the future, or you can create a recurring campaign that sends messages at set intervals. For more information, refer to Amazon Pinpoint Campaigns.

This Guidance includes a sample dataset of personalized car searches that is used to train the machine learning (ML) model. It also includes a demo that shows how to use ML to make product recommendations and automatically update your endpoints and segments. You can build upon this architecture for a variety of use cases.

 

Predictive User Engagement

Version 1.0
Last updated: 11/2019
Author: AWS

Additional resources

Did this Guidance help you?
Provide feedback 

Features

Automation

Build an architecture that automatically updates Amazon Pinpoint endpoints with predictive recommendations from Amazon Personalize.

Demo

This Guidance includes a sample dataset of personalized car searches and a demo walkthrough you can use to demonstrate functionality.
Build icon
Deploy a Solution yourself

Browse our library of AWS Solutions Implementations to get answers to common architectural problems.

Learn more 
Find an APN partner
Find an APN Partner

Find AWS certified consulting and technology partners to help you get started.

Learn more 
Explore icon
Explore Solutions Consulting Offers

Browse our portfolio of Consulting Offers to get AWS-vetted help with solution deployment.

Learn more