Search & Merchandising

Sort Options

Configure shopper-facing product order controls such as Featured, Best Sellers, Most Popular, Newest, and Price.

Sort options let a storefront expose familiar ordering controls while keeping the ranking logic consistent across platforms. Sort options can be used on product listing pages, collection pages, search results, and other catalog surfaces that request products from PSYKHE AI.

Sort Patterns

Sort patternBehavior
FeaturedThe default PSYKHE AI personalized product order for the surface.
Best SellersProducts with the strongest recent commercial engagement.
Most PopularProducts with broad recent engagement across shoppers.
Attribute-basedProducts ordered by a configured product or variant attribute. Common examples include newest first, price low to high, price high to low, inventory, vendor, product type, or a custom metafield.
CuratedA merchant or operator-managed order, such as a campaign or editorial collection.

The visible labels and display order are configurable per storefront. A store can expose attribute sorts as familiar shopper-facing labels such as Newest, Price: Low to High, or Price: High to Low, and can expose activity windows with labels such as Best Sellers - 30 Days.

Activity-Based Sorts

Best Sellers and Most Popular use recent storefront activity to order products. They are not personalized recommendations: every shopper sees the same order for the same catalog, filters, and category until the sort updates.

Depending on the available tracking, these sorts can use product views, add-to-cart events, and purchases, with stronger commercial events weighted above lightweight browsing events. They are not all-time sales reports.

Inputs Used

These sort options use storefront events and catalog freshness when they are available:

SignalRole in the ranking
Product viewBaseline interest signal.
Add to cartStrong intent signal.
Purchase / transactionStrongest commercial signal when tracked.
Product publication dateFreshness fallback for products with little or no engagement.

Storefront events are counted inside the configured window, and newer interactions have more influence than older interactions in the same window.

Product list views and impressions are useful for analytics and evaluation, but they should not be used as positive popularity signals by themselves because they mostly measure where products were already shown.

Products without recent events are still eligible. They are ordered after products with engagement using product freshness. This keeps the sort complete even for new catalogs, quiet environments, or products that have not yet gathered traffic.

Time Windows

Best Sellers and Most Popular can be configured over these time windows:

WindowWhat it measures
7 daysActivity from the last 7 days.
14 daysActivity from the last 14 days.
30 daysActivity from the last 30 days.
45 daysActivity from the last 45 days.
60 daysActivity from the last 60 days.

The right window depends on traffic volume, catalog size, buying cycle, and how often the merchant wants the sort to change. Choose a shorter window when recent movement matters and there is enough event volume. Choose a longer window when orders or engagement are sparse and the sort needs a more stable sample.

Different windows can be exposed as separate sort options when that helps the shopper or merchandising team, or kept implicit behind a simple label such as Best Sellers.

On this page