Use cases - Maintaining Personalized Experiences with Machine Learning

Use cases

You can use Amazon Personalize to:

Drive better user engagement

Use personalized user recommendations to help drive better engagement and conversion (USER_PERSONALIZATION recipes).

Improve customer search experience

Personalized curated lists or search results for your users improve the customer experience and increase customer loyalty and engagement. Amazon Personalize helps create a personalized list by re-ranking a collection of input items based on predicted interest level for a given user (PERSONALIZED_RANKING recipes).

Keep users engaged with trending or popular items

If your customers highly value what other users are interacting with, Amazon Personalize can help recommend trending or popular items. Common uses include recommending viral social media content, breaking news articles, or recent sports videos (POPULAR_ITEMS recipes).

Increase user conversion by recommending similar items

Recommending similar items can help your customers discover items and can increase user conversion rate. Amazon Personalize can help recommend similar items, such as items frequently bought together or movies that other users have also watched (RELATED_ITEMS recipes).

Targeted marketing campaigns through user segments

Getting user segments can help you create advanced marketing campaigns that promote different items to different user segments based on the likelihood that they will take an action. Amazon Personalize can help generate segments of users based on item input data, such as users who will most likely interact with items with a certain attribute (USER_SEGMENTATION recipes).

To increase the adoption for Amazon Personalize, you can use the Maintaining Personalized Experiences with Machine Learning solution to:

Create automated Amazon Personalize recommendation pipelines

Select an appropriate recipe per your use case and create automated pipelines as simply as providing json files and data in a S3 bucket.

Run retraining and inferencing via scheduled cron jobs

As a part of the json configuration, schedules can be setup to retrain the model fully/update it at a desired frequency. Campaigns, batch inferencing, etc. can also be made to use the updated models.

Advanced error handling and status updates

The solution provides error messages and job completion statuses via an SNS topic and email notification. This simplifies error handling and status updates and you can receive completed statuses for scheduled jobs, and in case of failures, an AWS XRay link to debug the issue.