Why not index them all
Duplicated content is the smaller risk. The larger one is dilution: a crawler spends its time on thousands of filter addresses and less on your products and categories. A page that exists in a hundred cuts also competes with itself.
What deserves a page
A combination that people search for as a phrase: blue hoodies, women’s waterproof jackets. It has volume, a clear meaning, and enough products. The signs that it deserves a page are the same as for a category: you could write a title and a sentence of introduction for it without inventing anything.
What Siftwren does
- every filtered page gets
noindex, followfrom the start, so links through it still count; - robots.txt keeps crawlers away from the filter parameters;
- the plain shop, category and product pages are untouched;
- with Pro, chosen combinations become pages with their own title, description, heading and introduction, are indexable, and are listed in your sitemap. See the landing pages guide.
A practical order of work
- Leave the defaults on and let the filters do their job for shoppers.
- Look in your search console after a few weeks. Filter addresses should appear as excluded by noindex or blocked by robots.txt, not as indexed.
- Find the two or three combinations your customers already ask for, and turn only those into pages.
Your categories remain the main landing pages. A filter set makes them easier to use; it does not replace them. Pro is for the shop that has gone through the three steps and wants the third.