LinkedIn’s content retrieval system enforces a hard 30-day window: after that, even your best-performing post is removed from the index entirely, making consistent publishing a structural requirement rather than a best practice.
Something strange happens to LinkedIn posts around day 31. A post that was still generating comments, still getting shared, still doing real work for the author’s visibility just stops. Not a slow fade. A hard cutoff. One day it’s in the system. The next day it’s gone.
This isn’t algorithmic decay. It’s infrastructure.
FishDB and the hard window
LinkedIn’s connection-based content retrieval runs through a system called FishDB, a Rust-based engine that handles posts from your connections, followed creators, and companies you follow. FishDB enforces a hard 30-day content window. After 30 days, a post cannot be retrieved through this path regardless of how relevant it is, how much engagement it earned, or how perfectly it matches a member’s interests.
This isn’t a relevance score declining over time. It’s a wall. The content is removed from the retrieval index entirely. The 30-day window is a design choice that keeps the index fast and manageable at the scale of over a billion members.
For content from people outside your network, the rules are different. The Causal LLM handles out-of-network retrieval through semantic embedding matching, and LinkedIn hasn’t published a comparable hard window for that path. But for the majority of reach that comes through existing connections and followers, 30 days is the ceiling.
What this means for “evergreen” content
B2B content marketers love the idea of evergreen content. Write something once, let it compound. And within the 30-day window, that logic holds. LinkedIn’s retrieval system prioritizes relevance over recency, so a strong post published two weeks ago can still surface above a mediocre post published this morning. Quality beats timing within the window.
But the window has a wall at the end of it. The most insightful framework breakdown you’ve ever written, the post that generated 200 comments and reshaped how people in your niche think about a problem, hits the same 30-day cliff as a throwaway observation you posted half-asleep.
This reframes what “evergreen” means on LinkedIn. It doesn’t mean “write once, benefit forever.” It means “write something good enough to keep working for 30 days instead of 3.” That’s still a meaningful advantage. Most posts burn through their reach in the first 48 hours. A post that’s still generating engagement in week three is outperforming 95% of content on the platform. But the compounding stops at day 30 whether you want it to or not.
The consistency tax most people don’t realize they’re paying
The 30-day window turns posting consistency from vague best practice into a retrieval requirement.
If you post three times a week, you’ve always got roughly 12-13 posts in the active retrieval index. Each one is a potential match when a connection opens their feed and the system selects candidates. Skip two weeks, and you’ve got maybe 3-4 posts left in the window. Skip three weeks, and you’re down to almost nothing retrievable. You haven’t been penalized. You’ve just disappeared from the candidate pool.
This is different from an algorithm punishing inactivity. The system doesn’t demote you or flag your account. There’s no penalty to recover from. You simply don’t have content available for retrieval. It’s the difference between being sent to the back of the line and not being in the building.
For B2B content marketers managing a company’s LinkedIn presence alongside ten other priorities, this has practical implications. A two-week vacation or a month focused on a product launch doesn’t just reduce your visibility temporarily. It creates a gap in your retrievable content inventory. The system can’t show what doesn’t exist in the index.
The resharing question
This raises an obvious tactical question: can you reshare older content to reset the clock? The architecture suggests yes, with caveats. A reshared post creates a new piece of content with its own embedding and its own 30-day window. But LinkedIn’s finalization layer applies impression discounting, meaning previously-seen content is heavily down-ranked. If most of your audience already saw the original, the reshare starts at a disadvantage.
The better approach is to treat the 30-day window as a production constraint. Your best ideas have a limited shelf life on the platform, which means they need to be published in a cycle. Not literally reposting the same content, but revisiting your core themes from different angles within each 30-day window so there’s always something strong and current in the index.
The deeper strategic question
The 30-day cliff reveals a fundamental tension in B2B content strategy on LinkedIn. The platform rewards depth and substance (the retrieval system matches topically rich content with the right audience, the ranking model learns from authentic engagement), but it imposes a hard expiration date on every piece of content you create.
This means LinkedIn isn’t a library. It’s closer to a newspaper. Your best work doesn’t accumulate. It rotates. The strategic response isn’t to fight this or complain about it. It’s to build a content operation that treats the 30-day window as a design constraint, the way a newspaper treats the daily print deadline. You need a steady supply of strong material because last month’s edition is already gone.
The authors who build real, sustained visibility on LinkedIn aren’t the ones who occasionally publish something brilliant. They are the ones who always have something worth retrieving in the index. That’s less romantic than the “one great post can change everything” narrative. But it’s how the infrastructure actually works.