The Nexus category tree lost two of its levels

The category menu used to fan out into four levels of nesting. It now fans out into two. Bookmarks pointed at the deeper pages generally do not resolve.

The listings that lived at the deep nodes are still on the market. What changed is the address of their category page, and the fact that a lot of intermediate pages that used to exist as landing points have simply been removed.

How it read before

The old tree looked like a librarians catalogue. A top category held subcategories. Each subcategory held a type. Each type held a variant. That gave you four clickable levels between the root and the listing grid. Each level had its own landing page with a filtered result set. You could bookmark any of them and land on the same shape of page as before.

A lot of those variant pages had very few listings. Some had none. But they were reachable, and a slow reader could descend the tree step by step, seeing the count fall at each hop, until only the products they had in mind were left.

How it reads now

The tree is two levels deep. A category holds types. The intermediate subcategory and variant levels are gone. What used to be a fully-clickable set of nested pages is now a two-hop navigation. Below the type page you are inside the search grid, filtered by the type, and any further narrowing is done inside the collapsed filter sidebar.

The mega-menu dropdown from the top navigation has been trimmed to match. It shows only what still exists. Category type is one click. Nothing further nests.

Some of the old deeper URLs redirect to the type page above them, which reads a little strange because a reader looking for a narrow variant lands on a broader shelf. Others do not redirect at all and return the standard not-found screen. Which URLs still redirect and which do not is not consistent across the tree, so a reader with a book of old bookmarks will find some working and some not.

Why this probably shifted

The first reading is that most of the deep pages carried nothing. An empty landing page is worse than no page at all. Deleting the empty nodes and folding the surviving ones back up the tree simplifies both the map and the moderation of it.

A second reading is that the tree was doing work that search now does better. If a reader types a term and lands in the results panel, they no longer need a five-page taxonomic descent to reach the same set. The tree becomes a shortcut for browsers, not a filing system for readers who know what they want.

A third reading is fewer indexable pages. That matters less on a market that is not indexed on the clear web, but it still reduces the surface a copy would need to reproduce to look convincing.

What to change on your side

• Assume any bookmark to a deep category page will 404 now. Rebuild them as filtered search URLs. The filter state travels with the URL as noted in the collapsed sidebar entry.
• Do not rely on the tree for narrowing. Use search plus filters. The tree gets you to the type; the filters do the rest.
• Do not read a missing subcategory as the market removing a product line. It is a shape change, not an inventory change. Related note: scanning listings fast.
• If the type page returns nothing on load, try flipping ships-from off. Some type pages now inherit a default filter that a returning reader will not expect.
• For advanced habits, check the pattern shift in how readers read reviews, which sits inside listings and did not depend on the tree.

What this entry is not claiming

This entry does not claim listings were removed. It does not claim any category is banned that was not banned before. It does not claim vendors were forced to re-tag their products. Only the tree that navigates to them has been shallowed.