How do you control faceted navigation without blocking useful ecommerce landing pages?

Separate navigation functionality from indexable landing-page strategy. Users can have many filters without every filter combination becoming a search landing page.

Start with a facet inventory. For each filter, list possible values, generated URL patterns, approximate URL count, and whether the combination represents a stable search concept.

Then classify combinations into three groups:

1. Indexable landing pages: combinations with distinct demand, sufficient inventory, stable products, unique descriptive value, and a clear place in the category hierarchy. These should have self-referencing canonicals, indexable robots directives, unique titles/headings, and internal links from relevant navigation or hubs.

2. Crawlable but non-indexable states: useful for users but not valuable as search results. Depending on architecture, use noindex and carefully consider whether links should remain crawlable for discovery.

3. Crawl-restricted parameter explosions: sort orders, session parameters, duplicate display modes, and near-infinite combinations that provide no discovery value. Prevent them from creating an uncontrolled crawl space through application rules, parameter handling, and link generation.

Do not canonical every filtered page to the parent category as a substitute for architecture. A canonical is a consolidation signal, not a crawl-control system.

Monitor logs and Search Console for parameter crawling, duplicate indexing, and valuable long-tail pages. The objective is to expose a curated set of useful facet landing pages while keeping the combinatorial space from consuming crawl resources.