What are Product Recommendations?
Product Recommendations use Adobe AI and machine learning trained on aggregated shopper behavior and catalog data to deliver personalized recommendations on Adobe Commerce storefronts. This overview covers service constraints (including HIPAA), data privacy, recommendation unit placement, storefront implementation options, product relationship integration, and catalog data retention.
Data handling and privacy
Data collection for Product Recommendations does not include any personally identifiable information (PII). All user identifiers such as cookie IDs and IP addresses are strictly anonymized. To learn more, see the Adobe Privacy Policy.
For more information about data syncing, see the Data Management Dashboard.
Where recommendations appear
Recommendations appear on the storefront as units with labels, such as “Customers who viewed this product also viewed.” You can create, manage, and deploy recommendations across your store views from the Adobe Commerce Admin. If your Commerce project uses the Adobe Commerce Optimizer Connector, you create, manage, and deploy recommendations through Adobe Commerce Optimizer.
Storefront implementations
Choose the documentation that matches your storefront:
- PWA Studio — PWA documentation
- Custom frontends (for example, React or Vue.js) — Integrate Product Recommendations in a headless storefront
- Commerce Edge Delivery Services (EDS) — Adobe Commerce Storefront documentation for EDS
Product recommendations versus product relationships
Given the ever-changing complexities of online shopping, what works best for your storefront is often a combination of multiple key technologies. Using both Product Recommendations and Product Relationships gives you more flexibility when promoting products. To automate your recommendations at scale, you can leverage Product Recommendations powered by Adobe AI. Then, you can leverage Related Product Rules when you must manually intervene and ensure that a specific recommendation is being made to a target shopper segment, or when certain business goals must be met.
Product recommendations allow you to:
- Choose from nine distinct intelligent recommendation types based on the following areas: shopper-based, item-based, popularity-based, trending, and similarity-based
- Use behavioral data to personalize recommendations throughout the shopper’s storefront journey
- Measure key metrics relevant to each recommendation to help you understand the impact of your recommendations
Product recommendations demo
To learn about Product Recommendations, watch this video:
Catalog data retention policy
The Product Recommendations service depends on catalog data that stays in sync with your Adobe Commerce environment. Inactive catalogs or environments that stop querying that data can become inactive, which affects what the service returns until you reactivate.
If you do not submit a query for the catalog data in your testing environment for 90 consecutive days, the catalog data is set to inactive mode and no data is returned for any query. The 90-day rule does not affect catalog data in your production environment.
If your environment has an empty catalog 45 days after being created, the catalog data is set to inactive mode and no data is returned for any query. This applies to both production and testing environments.
Reactivate catalog data
To restore catalog data after it becomes inactive, submit a support request with the title “Reactivate Product Recommendations” and include the environment IDs. Catalog data should be restored within two hours.