Streaming home pages present rows with unusually narrow labels rather than broad genres. Those labels are generated from a tagging system, not written by an editor for each user.
Catalogs are tagged far below genre level
Every title in a large catalog is described with many attributes: tone, setting, era, pacing, subject matter, the kind of protagonist, and more.
Human taggers apply these using a controlled vocabulary so the same concept is recorded identically across thousands of titles. Consistency is the entire point.
Genre alone is far too coarse for recommendation. Two crime dramas can share a genre and share almost nothing an actual viewer cares about.
Row labels are assembled, not authored
A row title is typically built by combining tags into a template, then filling it with whatever titles satisfy the combination for that viewer.
Because the combination is chosen by a system optimizing for engagement, it can join attributes no editor would pair. The oddness is the seam showing.
The labels are also a form of explanation. Telling a viewer why a row exists makes the suggestions easier to accept than an unlabeled grid.
The home page is assembled per session
Rows are selected and ordered for each visit based on viewing history, time of day, device, and what the service currently wants to surface.
Two people in the same household therefore see different pages, which is a large part of why profiles exist at all.
Position matters enormously. A title in the first row of the first screen gets attention that the same title cannot get four rows down.
Artwork is selected alongside the row
Services generate multiple images for a single title and choose among them based on what a given viewer has responded to previously.
The same show can appear with different artwork on different accounts, emphasizing a different character or a different mood.
This occasionally misleads. An image chosen to match a viewer's history may promise a tone the title does not sustain.
Narrow rows help small titles surface
A broad genre row will always be dominated by the biggest titles. A narrow row is the only place a modest title can plausibly appear first.
Services need that mechanism because a catalog full of unwatched titles is expensive. Licensing and production costs are paid regardless of views.
The strange labels are therefore doing economic work. They create shelf space where a title has a chance of being noticed at all.