Guide · LinkedIn data

The LinkedIn algorithm changed. 12,697 posts say how.

In May we pointed Claude Code at 6,820 posts and wrote down what made them go viral. Three months later, half of those rules had stopped working. So we ran it again — 37 creators, 12,697 posts, spring against summer, post by post.

10 steps · ~25 minClaude Code · ColdIQ MCP · 12,697 posts · 37 creators

What you will build

In May, we pointed Claude Code at 6,820 posts from the best creators in GTM, sales, AI and content and wrote down what made them go viral. Three months later, half of those rules stopped working. So we ran it again, on more creators and more posts, and compared spring to summer post by post.

You get two things here. Steps 01 to 03 are the method — the setup and the exact prompts, so you can run the same analysis on the creators of your own niche in an afternoon. Steps 04 to 10 are what we found, with every number.

How we measured it

37 creators who consistently go viral posting about AI, sales, GTM and content: founders, operators and full-time creators, from ~8K to 860K followers. 12,697 posts in the library.
Spring (1 Jan to 25 May 2026) against summer (26 May to 31 Aug 2026). September is left out — its posts are too recent for likes to be final.
Every post is scored against its own creator's spring median likes. 1.0 is a normal post for that creator, 2.0 is twice their usual — so an 860K account and an 8K account are compared fairly.
A breakout is a post that gets at least 2x its creator's usual likes. Lift is how often a type of post breaks out compared to the average post of the same period: 1.0 is average, 2.0 is twice as often.
Reposts and duplicates removed. J Hawkins (PostHog) is left out of the rates — he posts around 15 one-line memes a week, which would skew every length and format number on his own. His posts still appear in the decodes.

Every number came out of the data. Where a segment is thin, it is flagged as a small sample rather than quietly dropped.

Prerequisites

What you need before you start

01 · Claude Code

To run the analysis. npm i -g @anthropic-ai/claude-code

02 · A ColdIQ account

To pull the LinkedIn posts. One API key gives Claude Code access to 40+ data providers, LinkedIn posts included. Free tier: 300 credits on signup, no credit card. Starter $99/mo (7,000 credits), Pro $199/mo (15,000), Scale $499/mo (40,000).

03 · A list of creators

The LinkedIn profile URLs of the people who consistently go viral in your niche. We used 37. Twenty is enough to see the shape.

Unlock the full guide

Enter your email to read the rest of the tutorial.

Copy-paste code. Any email works. Unlocks in your browser, so you only do this once.

Step 01

Connect ColdIQ to Claude Code

Everything in this guide was built with Claude Code and the ColdIQ API. Sign up, pick a plan, then copy your API key from the ColdIQ marketplace. One key gives Claude Code access to 40+ data providers, LinkedIn posts included. Paste the line below in your terminal with your key, then restart Claude Code so it picks up the server.

Step 02

Pull the posts

Give Claude Code the list of creators who consistently go viral in your niche. Twenty is enough to see the shape; we used 37, from roughly 8K to 860K followers. Ask for one JSON file so the next step has a single source to work from.

Unlock above to read the full guide

Want us to build agents like this for you?

Want this funnel built for you?

This whole analysis was built with Claude Code and the ColdIQ API in an afternoon. ColdIQ builds AI-powered GTM systems for B2B teams every day — the content engine, the enrichment behind it, and the sequences that turn engagers into conversations. If you would rather have it wired for you than assemble it yourself, book a call.

Talk to us

Your GTM stack. One API key.

Stop managing 39 tools. Start shipping GTM from code.

Get Started Or try the playground first

Free to start. No credit card required.