What changed
Google now shows ordinary LinkedIn posts far more than it used to, and long-form LinkedIn articles far less.
- Ahrefs estimates that Google traffic to LinkedIn articles fell about 85% between March 2024 and January 2026, and the number of indexed article pages fell 92%. [Growtika]
- Over roughly the same period, its estimate of Google traffic to ordinary posts rose about 180%, from 2.5 million to 7 million visits a month. [Growtika]
Both are tool estimates for US search, so read them as direction, not exact counts.
AI tools tell a different story. In Semrush's study of 89,000 LinkedIn pages cited by ChatGPT Search, Google AI Mode and Perplexity, articles made up 50 to 66% of the LinkedIn content cited and posts 15 to 28%. [Semrush] Posts are winning in Google, while articles still matter for AI answers.
How Google reads a LinkedIn post
Google treats a public post like any other web page, and the first line does most of the work.
Product liability insurance for manufacturers is where most plant owners are underinsured. Here are the three gaps we see in almost every policy we review.
- The web address comes from roughly the first seven words of the post, and it never changes after you publish. [Toth] The live addresses we checked used about the first 40 to 45 characters.
- The title usually comes from the opening text. Sometimes LinkedIn writes its own headline from the post instead. [Toth]
- According to SEO practitioner Steve Toth, if a post opens with hashtags, they become its web address and title. [Toth]
Most plant owners find out they're underinsured the day after a recall.The address says "most-plant-owners-find-out". Google has nothing to match the search to.
Product liability insurance for manufacturers is where most plant owners are underinsured.Same hook, with the phrase in front of it. The phrase is now in the address, the title and the opening sentence.
This isn't a case against hooks. A good hook is what makes people stop and read. For a post you want found in search, the phrase goes first and the hook comes straight after it.
When to use this, and when not to
Most of your posts should still be written for the feed. Use the phrase-first approach for the few that do a search job. One a week alongside your usual posts is a sensible start.
A fast lane for newer websites?
If your own site is new or has few links, it can take a long time to reach page one for a contested search. A LinkedIn post sits on one of the most trusted domains on the web, so it can get there much sooner. In one study, posts that newly reached Google's top three were a median of 62 days old. [Toth]
- It works for specific questions, niche phrases and ideas you name, especially where Google already shows LinkedIn posts. Search the phrase in a private window. If posts are already on page one, a well-built post can join them.
- It rarely works for the most competitive searches, where big brands and directories hold page one and LinkedIn posts don't appear.
- It doesn't last on its own. In one study 37% of top-three posts were gone within a week. [Toth] Keep publishing on the phrase.
- The traffic isn't yours. Clicks land on LinkedIn, so each post needs to send people on to a page on your site.
- The platform can change the rules. Google traffic to LinkedIn articles fell about 85% in two years. [Growtika] Posts could go the same way.
Treat it as a bridge, not a replacement. Use posts to show up for specific searches now, and link every one back to your site so your own pages build authority for the long run.
What the studies say decides which posts rank
| What we looked at | Evidence | What the data shows |
|---|---|---|
| Links from other sites | Measured | In a study of 645 posts, 31% of posts with at least one link from another site ranked in Google's top 20, against 7% with none. 91% of the LinkedIn pages in that dataset had no links. [Indie Hackers] |
| The phrase in the opening lines | Small study | Of 60 top-ranking posts, about 60% contained the exact search phrase, and roughly half had it in the first two lines. [Toth] |
| Reactions and comments | Measured | Posts with 200+ reactions ranked 20% of the time, against 4.5% under 10. [Indie Hackers] Yet number-one posts had fewer reactions (median 37) than posts at two and three (median 118). [Toth] |
| Age and churn | Measured | The median ranking post was 300 days old, against 180 for posts that didn't rank. [Indie Hackers] In another study, 37% of top-three posts were gone within a week. [Toth] |
| Length and formatting | Mostly articles | Length, headings, lists, emoji and hook length didn't separate pages that ranked from pages that didn't. [Indie Hackers] The median ranking post was 167 words. [Toth] |
| Video | Small sample | After one week, 13 of 14 video posts were still ranking, against 4 of 8 text-only posts. [Toth] |
These are observational studies, mostly of US searches about SEO and B2B software. They show what ranking posts have in common, not what causes a post to rank.
Seven things to change in how you post
- Pick the phrase before you write. For a post meant to be found, a search a buyer would type or a name for an idea you own.
- Put the phrase in front of your hook. State the phrase, then hook in the same sentence or the next. Your first line is your page title.
- Answer early. Google shows the start of the post as its description. 150 to 300 words is plenty.
- Set the post to Anyone. LinkedIn says only posts set to Anyone may be crawled by search engines and AI tools. [LinkedIn Help]
- Keep hashtags out of the first line. Put them at the end, or leave them out.
- Link to your best posts from your own site. Links were one of the clearest signals in the data.
- Publish the idea more than once. Rankings churn. Come back to the phrase from a new angle every six to eight weeks.
Posts or articles?
Use both, for different jobs. Posts are where the Google traffic is. Articles are what AI tools cite most. But a LinkedIn article always points its canonical tag at itself, and you can't point it at your own site. [LinkedIn Help] Pasting a full blog post into an article creates a second copy competing with your original, so publish on your own site first and post a shorter, different take that links back.
What we're testing ourselves
Most of this evidence is American and some of it rests on small samples, so we're running the tests on our own posts for eight weeks: phrase first against hook first, hashtags at the start against the end, Anyone against connections only, and linked posts against unlinked ones. We'll publish the results here, including anything that contradicts the studies.
Sources
- Growtika, LinkedIn Pulse research, February 2026. Ahrefs estimates, US.
- Semrush, We Analyzed 89K LinkedIn URLs Cited in AI Search, 10 March 2026.
- Indie Hackers, LinkedIn SEO, 5 steps from 1,260 searches and 9,805 URLs, 11 August 2026. US desktop, English.
- Steve Toth, What 60 LinkedIn Posts Ranking in Google and AI Overviews Have in Common, 10 September 2026. US.
- LinkedIn Help, Availability of your content off LinkedIn, and SEO settings for articles.