If you generate recommendations outside Blueshift, you can still deliver them through Blueshift campaigns. Choose one method, or combine methods, based on where your recommendation data lives and how often it changes.
Pick a method
| Method | What you provide | How recommendations are supplied | Learn more |
|---|---|---|---|
| Feeds | A file or warehouse table mapping user → products or product → related products | Imported into Blueshift and stored as a recommendation feed | Details ↓ |
| Events | A custom event per user containing product IDs | Products are sent or imported into Blueshift through a custom event and updated using Save Products with Replace existing. | Details ↓ |
| External fetch | An external endpoint that returns dynamic content as JSON | Fetched from your system just before Blueshift sends the message | Details ↓ |
Catalog requirement
Product IDs must exist in your Blueshift catalog so recommendation blocks can resolve attributes such as title, image, and price.
How this connects to schemes
No matter how your recommendations enter Blueshift, you deliver them the same way: create a recommendation scheme, add a block that reads your recommendation data, and activate it in your message or template.
For a step-by-step guide to setting up a recommendation scheme, see Build a recommendation scheme.
Feeds
Use feeds when you have a precomputed list from your data science, BI, or partner systems. Blueshift stores the mapping and serves it through feed-based recommendation blocks.
user@example.com → SKU_001, SKU_002, SKU_003
Set up feeds in Customer AI → Recommendations → Feeds → + Content Feed.
- Recommendation feeds import from file (CSV, S3, SFTP)
- Recommendation feeds import from data warehouses (Snowflake, Databricks, BigQuery)
Next step: use the feed in a scheme
After importing recommendations using a CSV file or a data warehouse, create a recommendation block and select a recipe to use your feed data. See Build a recommendation scheme for the list of recipes and step-by-step instructions.
Events
Use events to send your own recommended items to Blueshift as a custom event, then use them in a recommendation block.
- Create a custom event, for example
user_fitment_products. - Enable Save Products for the event.
- For recommendation use cases, select Replace existing as the product saving mode.
- Use a recipe that reads from that event inside a recommendation scheme.
To learn more about event properties and custom event configuration, see Events.
To send or import event data into Blueshift, see Import events data.
To configure the event input in a recommendation block, see Build a recommendation scheme.
External fetch
Use external fetch when your recommendation data is available from an external endpoint and you want Blueshift to retrieve it at send time.
External fetch requires setup by the Blueshift team. Contact your CSM or support@blueshift.com to discuss your requirements.
For background information about how external fetch works, see External fetch.
What to do next
- Choose the method that matches your data source.
- Configure feed import, event delivery, or event import. For external fetch, contact your CSM or support.
- Build a scheme that reads your recommendation data: Build a recommendation scheme.
- Add the recommendation to your messaging template: Activate recommendations.
Small fixed list?
If you only need a small, static set of items, use a curated or manual items block and enter product IDs directly in the block Inputs.
Comments
0 comments