When we first wrote about LinkedIn's war on engagement pods, the platform was still issuing warnings. Pods were annoying, cringe, and occasionally throttled — but they worked. Founders ran them. Ghostwriters sold them. Entire "LinkedIn growth" agencies were quietly built on a WhatsApp group and a Google Sheet.
That era is over. In early 2026, LinkedIn ripped out its entire ranking system and replaced it with a 150-billion-parameter AI model called 360Brew. Pod detection now runs at 97% accuracy. Accounts that trip it watch their reach collapse from roughly 8,500 impressions per post to around 340 — overnight — and recovery takes 60 to 90 days of fully compliant behaviour. Pods are now an explicit LinkedIn ToS violation. Generic "Great post!" comments, "Agree?" hooks, and rapid-fire like patterns are all actively downgraded.
If you're running, buying into, or building content strategy around engagement pods in 2026, you are burning distribution you cannot get back. Here is what the new system actually rewards.
The Crackdown Is Complete
LinkedIn has been chipping at engagement pods for three years. In 2026 it finished the job. Pod detection now runs at 97% accuracy, driven by reciprocal-engagement pattern analysis — if the same small cluster of accounts repeatedly likes and comments on each other's posts within minutes of publishing, the system flags every participant.
The consequence isn't a warning, a notification, or a nudge. It's algorithmic death. Flagged accounts are quietly demoted for months. According to ConnectSafely's April 2026 reporting, operators of the largest pod networks have watched entire client rosters go dark simultaneously — a pattern that only makes sense if LinkedIn is now identifying the pod itself and penalising every node in it.
Participating in a pod in 2026 isn't a growth hack. It's a permanent scar on your account.
How 360Brew Changed the Rules
The old LinkedIn algorithm was a patchwork of hand-written rules, engagement weights, and topic-classification models bolted together over a decade. 360Brew replaces the whole thing with a single large model trained directly on what LinkedIn users value: meaningful interactions, dwell time, and genuine professional signal.
Two shifts matter most:
- Pattern detection is holistic. 360Brew doesn't just count likes — it looks at who's liking, how fast, from what network, with what comment quality. Coordinated behaviour is obvious to a 150B-parameter model in a way it never was to a rules engine.
- Content quality is scored continuously. Generic content — the sort of recycled "5 lessons from my first million" posts that used to dominate feeds — now gets up to 50% of its reach automatically suppressed. The model can tell the difference between a template and a thought.
The net effect: the surface area for gaming LinkedIn has collapsed. The surface area for saying something genuinely useful has expanded dramatically.
What Pod Users Are Seeing: The 8,500 → 340 Collapse
The numbers being reported by flagged accounts are brutal and consistent. Before flagging: typical posts reaching 8,000 to 9,000 impressions with 200-plus reactions. After flagging: the same account, same content style, same posting cadence — 340 impressions. A 96% collapse.
Three things make this worse than a normal algorithmic dip:
- It's account-level, not post-level. Every post you publish while flagged is suppressed, not just the one that tripped the detector.
- It's sticky. Recovery requires 60 to 90 days of fully compliant behaviour — no pod activity, no automation, no reciprocal-engagement patterns — before reach begins to normalise.
- There's no appeal UI. LinkedIn does not tell you you've been flagged. You simply stop being seen.
For agencies and ghostwriters running pods as a service, this is existential. For individual operators, it's a six-figure mistake: a year of audience building, wiped out in a week, with a three-month rebuild on the other side.
What Works Instead: The Depth Score Era
360Brew's new primary ranking signal is what LinkBoost and others have started calling the Depth Score — a composite of dwell time, scroll-stop rate, comment quality, and save/share actions. Likes barely move the needle anymore. Depth does.
What that means in practice:
- Write posts people actually finish reading. A 120-word post that holds attention for 18 seconds outperforms a 1,200-word post skimmed in four.
- Earn real comments, not reactions. A handful of multi-sentence comments from people in your target audience will beat 200 generic likes every time.
- Optimise for saves. Save rate is one of the strongest Depth Score inputs, because saving a post is a costly action — users only do it when the content is genuinely useful.
The irony is obvious: the behaviour pods were designed to fake — meaningful professional engagement — is exactly what 360Brew now measures with enough precision that faking it is worse than not trying.
The 60-Minute Window That Decides Reach
Under 360Brew, the first 60 minutes after a post goes live determines almost everything about its eventual reach. This is the window in which the model decides whether to push the post into second- and third-degree networks or quietly bury it.
What the model is looking for in that first hour:
- Comments with substance (not "Great post!")
- Dwell time above 8 seconds
- Saves and shares
- Engagement from accounts outside your immediate pod of regular commenters
What kills a post in that first hour:
- A flurry of likes from the same ten accounts that engage with everything you publish
- One-word comments
- Reciprocal engagement patterns (you commented on theirs 20 minutes ago, they're commenting on yours now)
This is why pods became suicidal. The exact behaviour pods were built to generate — fast, clustered, reciprocal engagement in the first hour — is the single clearest signal 360Brew uses to demote a post.
Text vs Video vs Carousel in 2026
The content-format hierarchy has inverted. For most of 2024 and 2025, long-form text posts with a "hook → story → lesson → CTA" structure ruled LinkedIn. In 2026:
- Micro content wins. Sub-150-word posts, native carousels, and short native video are outperforming long-form text across almost every B2B vertical.
- Video still wins raw reach. Native video posts get pushed hardest by 360Brew because dwell time compounds per second watched.
- Text wins lead quality. High-Depth-Score text posts — short, sharp, genuinely useful — convert to inbound DMs and qualified pipeline at rates video rarely matches.
- Carousels are the sleeper format. They combine high dwell time (users swipe through) with high save rate (users bookmark the whole deck), which is the exact Depth Score profile 360Brew rewards.
The playbook: lead with carousels and short native video for top-of-funnel reach, follow with tight text posts for pipeline.
What This Means for Marketing Teams
The strategic implication of all this is bigger than "stop using pods."
LinkedIn just told every B2B marketing team that their organic distribution model — the one built on personal brand ghostwriting, scheduled batches, and pod-juiced launches — no longer works. The new model requires genuinely useful content, posted with precise timing, optimised for the first-60-minute Depth Score window, across multiple formats, by real humans whose networks have real overlap with buyers.
That is not a job you can do with a spreadsheet, a freelance ghostwriter, and a Buffer subscription. It requires:
- Content generation that can produce on-brand, genuinely useful posts at volume
- Scheduling that hits your audience's peak-engagement windows, not yours
- Multi-format output (carousel, video, micro-text) from a single brief
- Performance analytics that track Depth Score proxies, not vanity likes
- Real-time distribution that can capitalise on the 60-minute window before it closes
Most marketing teams have none of this stitched together. Which is exactly why we built Anjin.
Anjin: The Marketing Operating System for the Depth Score Era
If 360Brew is what AI did to LinkedIn's ranking layer, Anjin is what AI does to your marketing operations layer.
Anjin is the Marketing Operating System — a single platform that plans, generates, distributes, and measures your content across channels. For LinkedIn specifically, that means:
- On-brand micro posts, carousels, and short-video scripts generated from a single brief, tuned to your voice.
- Scheduling built around your audience's first-60-minute window, not yours.
- Depth Score proxies tracked automatically — dwell, saves, comment quality — so you know what's actually earning reach.
- A content engine that ships campaigns the same day a news moment breaks, without a pod, a ghostwriter, or a £8–15k/month retainer.
This is how you run LinkedIn distribution in a 360Brew world: real content, real timing, real signal — at the cadence a single operator running Anjin can sustain indefinitely.
Sources: ConnectSafely, LinkBoost — Depth Score, LinkBoost — Pods 2026, Linkmate, GrowLeads, Bang Marketing.
