Introduction
Our Upsells API returns a ranked list of recommended products based on the content of a customer's basket. It is designed to increase average order value and basket size by surfacing relevant add-on items at the point of ordering, without requiring manual configuration of recommendation logic by the ordering platform.
How it works
Recommendations are generated using two signals;
- Analysis of the brand's historical order data - which products are genuinely bought together, beyond simple popularity (individual customer history isn't factored)
- An AI-based analysis of the brand menu - which products are comparable to or complementary to one another
A recommendation must be supported by both signals to be returned, which keeps suggestions specific to each brand's menu and the behaviour of their customer base.
The response will also filter on 'live menu' content, meaning that only products currently online and not snoozed at the specific location are recommended.
Upsell Overrides
Brands can control the recommendations via Upsell Overrides, which allow specific products to be "pinned" so they are always suggested at a given location. This aids in selling new promotions or seasonal items which wouldn't factor in order history rankings.
Overrides work at the individual product level (not category), are merged ahead of the automatic suggestions rather than replacing them, and duplicates are removed automatically.
Upsells Workflow
See below for a complete Upsells workflow diagram;
FAQ
See link below for some frequently asked questions;