Product catalog & recommendations

"Customers who bought X also bought Y" — inside your own messages, from your own data, with no external recommendation service.

1. Load the catalog

curl -X POST https://api.wezend.com/v1/catalog/products \
  -H "X-API-Key: $WEZEND_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{ "products": [
    { "sku": "TSHIRT-BLK-M", "name": "Black tee (M)", "price": 29.0, "currency": "EUR",
      "url": "https://shop.example.com/p/tshirt-blk", "image_url": "https://…/tshirt.jpg", "category": "apparel" }
  ] }'

Bulk import accepts your full feed; re-posting a SKU updates it. Most shops sync nightly from their commerce platform.

2. Feed it behaviour

The model learns from the events you already track — product_viewed, added_to_cart, order_completed (see Tracking events). Rebuild the affinity model on your schedule:

curl -X POST https://api.wezend.com/v1/catalog/affinity/rebuild -H "X-API-Key: $WEZEND_API_KEY"

It computes co-occurrence: which products appear together in real baskets and browsing sessions.

3. Use recommendations

Per contact:

curl "https://api.wezend.com/v1/catalog/recommendations/c_123?limit=4" -H "X-API-Key: $WEZEND_API_KEY"

In the email builder, add a product grid block set to "personalised" and each recipient gets their own four products at send time. In journeys, the post-purchase cross-sell ("order delivered → wait 7 days → send recommendations") is a template you can enable in one click.

Contacts with no history get graceful fallbacks: bestsellers in the categories they've browsed, or overall bestsellers — never an empty grid.