Recommendations

Recommendation models learn from the same view, click, add-to-cart and purchase events that feed personalization and analytics: which items tend to be viewed or bought in the same session predicts what belongs on a product page's “frequently bought together” rail or an empty cart's “you might like” panel.

Because they run off behavioural signal rather than manual curation, recommendations keep working as the catalogue turns over — a newly added product starts appearing once it accumulates its own signal, with no rule to write by hand.