Faceted navigation

What faceted navigation does

Faceted navigation is the filter panel in a listing: colour, size, price, brand, availability, material. It lets customers turn a large assortment into a short list within a few clicks. The difference from plain sorting is that several filters apply at once and constrain each other. Choosing red leaves only the sizes that actually exist in red.

For shops with more than a few hundred products the filter is not an extra, it is the main path to a choice. Anyone scrolling through 40 pages buys less often. This shows up in the bounce rate of the listing and in the share of orders that involved a filter interaction.

How facets are built

• Data source: Filter values come from product attributes or variant options. In Shopware the properties and the option values of variants feed the facets.
• Index: The values live in a search index, typically OpenSearch. Aggregations return the values together with the hit count per value.
• Filter logic: Within one facet values are combined with OR, across facets with AND. Red or blue, but always a shirt.
• Empty values: Values without hits must not be selectable, otherwise every click leads into a dead end.

SEO: which filter URLs may be visible

Filters generate a very large number of URLs. A shop with six facets easily reaches several thousand variants, and if all of them are indexable the list of relevant pages gets diluted. The usual rule is therefore that filter URLs with a single constraint receive noindex, follow. Combinations with real search demand become landing pages of their own, with their own copy, their own title and a self referencing canonical.

Consistency matters. Two technically identical URLs with a different order of parameters must share one canonical form, otherwise duplicates appear. A stable ordering of the parameters is enough.

Common mistakes

Three mistakes cost the most. First, empty result pages: if a filter value no longer matches the other filters, the shop must grey it out or explain it instead of showing an empty page. Second, nesting that is too deep, because hardly anyone works through more than two levels. Third, ignoring mobile: there the filter belongs in an off canvas with a visible hit count, otherwise the customer cannot tell what the click did.

What to measure

Three figures are enough: the share of sessions with a filter interaction, the bounce rate of the listing with and without filters and the conversion rate per filter path. If a filter is used often and still converts poorly, either the logic or the order is wrong. The most important facet belongs at the top, not the one with the most values.