Keep-everything is the wrong instinct when the pages you are keeping are thin, duplicate, or off-topic, because at that point the volume is dragging your quality signals down, wasting crawl on pages that earn none of it, and cannibalizing the queries your good pages should own. The pivot is simple: is the content you are hoarding a quality drag or a quality asset. When holding onto low-value pages is actively costing the site, the instinct to keep them all has stopped helping and started hurting.
The hoarding instinct rests on the idea that more content is always an asset, that every page is at worst neutral, so deleting is just throwing away potential. That holds only while the pages are genuinely useful or at least harmless. It breaks the moment the pages are low-value in a way that registers across the site. Thin pages lower the average quality of the indexed set. Duplicate pages split signals and confuse which version should rank. Off-topic pages stretch the site’s focus and absorb crawl attention that should go to your core material. None of those are neutral; each carries a cost, and keeping all of them means paying all of those costs.
So keep-everything is wrong precisely when the kept content tips from harmless to harmful. A large library of strong, distinct, on-topic pages is worth keeping in full, and there the instinct serves you. A library padded with thin rewrites, near-duplicates, and abandoned tangents is one where volume has become a liability, and there the instinct works against you. The question is never “how many pages do I have,” it is “are these pages, on balance, lifting the site or weighing it down.”
To decide, audit your pages against that pivot: separate the genuinely useful from the thin, duplicate, and off-topic, and judge whether the low-value group is large enough to be hurting. Where it is, cut or consolidate those pages instead of defaulting to keeping them, and reserve keep-everything for the content that actually earns its place.