Product recommendations
Every product in your catalog can surface a list of products customers commonly bought alongside it — the "customers also bought" row you see on most storefronts.
Recommendations are computed from your marketplace's own order history. Nothing needs configuring, and no list needs maintaining by hand.
How the list is built
- Source data: your recent completed orders, over a rolling window.
- Ranking: products bought in quantity alongside the source product rank above ones that appeared with it only once.
- Size: a short list — enough for a "customers also bought" row, not a full page.
- Freshness: results are cached briefly, so a newly popular pairing appears within the hour rather than waiting for an overnight job.
Because the list is computed on demand rather than precomputed in a nightly batch, a product that went live this morning can already appear in it.
The window, list length and cache duration are defaults Traide sets and tunes over time. If your integration depends on an exact number of recommendations, request the count you need rather than relying on the default.
What buyers can and cannot see
Recommendations respect the same visibility rules as the rest of your catalog. A recommended product is only shown to a buyer if it is published, active, visible and available to them — the same test used on product listing and detail pages.
This means the list can be shorter than six, or empty:
- A newly launched marketplace has no order history to draw on.
- A product nobody has co-purchased yet has no pairings.
- Recommendations that resolve to unpublished or unavailable products are dropped rather than shown as dead links.
An empty recommendations row is normal early on and fills in as orders accumulate.
Showing them on your storefront
Recommendations are delivered through the API, on the product itself, for your storefront to render. If you use the Traide storefront, they are available to your product pages; if you have built your own, your developers can request them alongside the rest of the product's data.
Building against the API? See the developer guide.
Related
- Cross-sell products — letting several sellers offer the same product, which is a different feature despite the similar-sounding name.