How to build product suggestions in Bubble

The simplest valid way to build product suggestions in Bubble is to store enough data about each Product (like category, tags, price range), then show a repeating group that searches for Products filtered by similarity to the current one or to the user’s behavior. The core is: you run a Do a search for with constraints that match what “similar” means, and optionally adjust the ranking using Sort by or Advanced filter when needed.

Core Structure

You need a Product data type with fields that describe similarity: for example Category (text), Tags (list of texts), Price (number). When the user views a Product page, Bubble knows the Current Page Product, and you can use that to search for related items.

• Make sure Products have strong metadata. Bubble can’t guess similarity unless you store it.
• Keep fields simple: plain text, lists of texts, numbers. Bubble handles these well in searches.

Building the Suggestion Repeating Group

Add a Repeating Group and set its Data Source to a search that expresses similarity. Example search:

// In Repeating Group's Data Source:
Search for Products
Constraints:
Category = Current Page Product's Category
Price >= Current Page Product's Price - 20
Price <= Current Page Product's Price + 20
Sort by: created date (descending)

This shows products in the same category and close price range. It’s simple but works immediately. If you want tag‑based similarity:

// Add an Advanced filter AFTER the search
:filtered (Advanced):
This Product's Tags intersects with Current Page Product's Tags

Use Advanced filters only when you must, because they run on the user’s device. If performance matters, store a single field like Primary Tag or Category and filter with regular constraints instead.

Optional: Suggest Based on User Behavior

Create a data type View with fields User and Product. Each time a user views a product, create a View. Then show suggestions by searching for Products whose Category matches the Category of the user’s most-viewed products.

• Search Views: filter by Current User.
• Pick the most common Category.
• Search Products with that Category.

This is simple to build and doesn’t require machine learning.