Recommendations are one of the oldest uses of AI in e-commerce, and still one of the best. Shoppers who click a recommendation spend more and return more often.
Three recommendations that work
- Bought together on the product page
- Recently viewed in the cart
- Back in stock by email

Start with good product data
Recommendations are only as good as the product data behind them. Clear titles, correct categories and up-to-date stock levels make a big difference.
Spend an hour fixing the top hundred products before you switch recommendations on. These products usually bring most of the sales.
Measure what matters
- Clicks on recommended products
- Average order value with and without recommendations
- Repeat purchases within ninety days
- Return rate for recommended items
Check these numbers every two weeks. If a recommendation type does not improve them, turn it off and test a different one.
Sassly reads your catalogue and orders, then serves recommendations through a small widget. Most stores see results in the first week.



