Faceted Search Design Template for PMs

Start with 5-8 facets on desktop. Research from the Baymard Institute shows that listing pages with more than 10 visible facets see lower filter adoption because the interface feels overwhelming. Use analytics to identify which facets have the highest selection rates and prioritize those. Less-used facets can live behind a "More filters" expansion.

Yes, if your search infrastructure supports it. Dynamic counts tell users how many results match each facet value, which prevents dead-end filtering (selecting a value that returns zero results). If dynamic counts are too slow, use the hybrid approach: update the currently active facet dynamically and cache others. For implementing the search backend, the [Search Relevance Template](/templates/search-relevance-template) covers index configuration.

OR is the standard for most facets. When a user checks "Enterprise" and "Startup" under Category, they want to see items that match either value. AND within a facet is rare and usually confusing. Cross-facet logic should be AND: "Category is Enterprise OR Startup" AND "Price is Free."

Add a search-within-facet input field. Show the top 5-10 values by count, with a "Show more" link or a text search to find specific values. Alphabetical sorting is useful for known-item search (the user knows the value name). Count-based sorting is better for discovery (the user wants to see what is popular).

Query parameters (`?category=x&price=100`) are more flexible and easier to extend. Path-based URLs (`/category/x/price/100`) look cleaner but are rigid. For faceted search with many possible filter combinations, query parameters are the standard choice. Path-based routing works better for fixed category hierarchies where you want SEO benefit from the URL structure. ---