LinkedIn's 2026 Algorithm Explained (Why Your Reach Dropped) — Transcript

Full transcript

1. 0:00LinkedIn just killed AI content. Most
2. 0:02people don't even realize that their
3. 0:04reach is getting cut in half. Let me
4. 0:06show you exactly what I mean. One of my
5. 0:08clients was literally getting 40 views
6. 0:09[music] per post. Then 28 days later,
7. 0:12when she started implementing everything
8. 0:13we're about to talk about in this video,
9. 0:15she hit 747,000
10. 0:17impressions. Same person, same effort,
11. 0:20just one simple thing that changed. By
12. 0:23the end of this video, you'll see
13. 0:25exactly what that one thing was. You'll
14. 0:26see her real results, and you'll
15. 0:28understand how LinkedIn's rebuilt
16. 0:30algorithm for 2026 changes how people
17. 0:32see you. And this is coming straight
18. 0:34from their own research papers. Plus,
19. 0:36the exact test that gets your post
20. 0:37pushed to thousands of people on
21. 0:38LinkedIn rather than getting buried.
22. 0:40Let's get into it. Here's what actually
23. 0:42happened. Researchers analyzed thousands
24. 0:44of LinkedIn posts just to realize that
25. 0:46over half of them were written by AI.
26. 0:48Since ChatGPT launched, that number rose
27. 0:51by 190%.
28. 0:53Everyone reached for the same tool, so
29. 0:55everyone sounds identical. And in 2026,
30. 0:58LinkedIn did something bigger than just
31. 1:00a tweak. They rebuilt the algorithm from
32. 1:02the ground up. The old system ran on
33. 1:04simple engagement signals, which is
34. 1:06exactly why the old tricks used to work.
35. 1:08That includes things like engagement
36. 1:09pods and also like engagement baiting
37. 1:11content, like lead magnets. [music] That
38. 1:13whole playbook is completely dead now,
39. 1:16because there's a completely new engine
40. 1:18underneath it. The new engine is an AI
41. 1:20model called 360 Brew. LinkedIn
42. 1:22published this research themselves, and
43. 1:24instead of counting likes, it actually
44. 1:25reads your post, your profile, and the
45. 1:27viewer's history like a person would.
46. 1:29And predicts one simple thing. Will a
47. 1:32real human engage with this? And yes, we
48. 1:35sit with the irony. LinkedIn's answer to
49. 1:37too much AI was using AI to filter out
50. 1:40other AI. Takes one to know one, I
51. 1:42guess. And here's the sneaky part.
52. 1:43Flagged posts don't get deleted. That
53. 1:45would be way too obvious. They just
54. 1:47quietly stop reaching anybody new. So,
55. 1:49the second your post sounds like it came
56. 1:51out of a prompt, like every other post
57. 1:53in the feed, the system knows, and it
58. 1:55stops showing you to out-of-network
59. 1:56connections. We already know what this
60. 1:58is. It's AI slop and you can smell it in
61. 2:00a second. The {m-dash} no human actually
62. 2:02types, the it's not this, it's that kind
63. 2:04of line. And the biggest giveaway of
64. 2:06all, the short incomplete sentences that
65. 2:08no human write. That's AI literally
66. 2:10trying to sound human and completely
67. 2:12missing. But here's the part that
68. 2:13actually kills your reach. These new
69. 2:14models read your post as language and
70. 2:17classify it from your opening. Your
71. 2:19first line or two literally decides who
72. 2:21it gets tested on. So, when you open
73. 2:23your post with a generic AI filler, you
74. 2:25don't just sound like a robot, you hand
75. 2:27the system nothing to work with. And it
76. 2:29sorts you into the wrong pile before
77. 2:30anyone reads a word. So, if the old
78. 2:32tricks are dead, what actually works?
79. 2:34Well, I didn't guess, I went to
80. 2:35LinkedIn's actual research papers, two
81. 2:37separate papers their engineers actually
82. 2:39published, and pulled out what the new
83. 2:41system is literally built to reward.
84. 2:42This is their language, not mine. In
85. 2:45their feed retrieval paper, LinkedIn
86. 2:46says their number one target is growing
87. 2:48something that they call professional
88. 2:50interactors, members who take real
89. 2:52action on a post, well time, comments,
90. 2:54reposts. But notice what's missing, a
91. 2:56like and the baby AI sounding kind of
92. 2:59comments. That doesn't move it at all.
93. 3:01In our experience, the single strongest
94. 3:03signal right now is a save. Think about
95. 3:06what it means when someone actually
96. 3:07saves a post. It's someone saying, "I
97. 3:09need to come back to this. It's so
98. 3:10valuable, I'll come back to it later."
99. 3:12That's the highest intent signal a human
100. 3:14can give you. So, we optimize every post
101. 3:16to be save-worthy and the reach follows.
102. 3:19And check this out, same paper, but this
103. 3:20one flips what everyone believes. The
104. 3:22new system leans hard on showing your
105. 3:24content to people that aren't connected
106. 3:26to you. And their own test data showed
107. 3:27the biggest gains went to the people
108. 3:29with fewest connections. Read that
109. 3:31again. Having a smaller account is now
110. 3:33an advantage because the new system is
111. 3:35hungry to fill those feeds with good
112. 3:38outside content. Reaching strangers
113. 3:40isn't the hard part anymore, it's
114. 3:41literally the entire point. And here's
115. 3:43something you can fix within the next 5
116. 3:45minutes, your headline. Because these
117. 3:47models read your profile as plain
118. 3:48language, your headline is one of the
119. 3:50strongest signals telling the algorithm
120. 3:52who to show you to. And sadly, most
121. 3:54people write it like a resume. Founder,
122. 3:57investor, speaker, advisor. That's four
123. 3:59job titles and it tells the model
124. 4:01nothing. If anything, it confuses it so
125. 4:03it can't match you to anyone and it just
126. 4:05stops showing you. And if you have a
127. 4:07headline like that, I'm sure you're
128. 4:08seeing 200 impressions, 400 impressions,
129. 4:10very low impression count. Now, watch
130. 4:12the fix. I help seed stage founders turn
131. 4:15LinkedIn content into book demos. Same
132. 4:17person, but now the model knows exactly
133. 4:19who you're for and who to put you in
134. 4:21front of. Your headline is not just
135. 4:22about a job title. It's literally
136. 4:24targeting the people you want to be seen
137. 4:25by. So, rewrite it to name who you help
138. 4:27and the outcome you get them. And now,
139. 4:29you've added the algorithm a map. Your
140. 4:31content works the same way. Go specific
141. 4:33and the models show you to the right
142. 4:34people's eyes. Stay vague and it shows
143. 4:36you to no one. So, line it up. The
144. 4:38system wants real interactions from
145. 4:40strangers on specific content it can
146. 4:42actually read. Now, picture the average
147. 4:44AI post. Generic, stuck inside your own
148. 4:46network, lucky to get maybe a like from
149. 4:49your friend, lucky to maybe get a like
150. 4:50from your coworker. It just completely
151. 4:52fails the new algorithm test. That's not
152. 4:54bad luck, that's just by design. So, to
153. 4:56actually win with this algorithm, you
154. 4:58need to do two things at once. Sound
155. 5:00human enough to clear the filter and
156. 5:01earn the real interactions the system
157. 5:03rewards. I call this the human test.
158. 5:05There's three signals. Signal one, an
159. 5:07original point of view. Not here are
160. 5:09five productivity [music]
161. 5:10tips, but the productivity advice I
162. 5:12followed for two years that quietly
163. 5:13killed my business. A real take or a
164. 5:16real story is what makes someone engage
165. 5:17for real. Signal two is real evidence.
166. 5:20Screenshots, real numbers, names, dates,
167. 5:23etc. Proof is what makes people save
168. 5:25your post and tag a coworker. Signal
169. 5:27three, a human voice, a real story, a
170. 5:30real rhythm. No em dashes. This is how
171. 5:32you sound like a human, not just like a
172. 5:34prompt. And that's how you clear the AI
173. 5:36filter in the first place. Do all three
174. 5:37and you stop getting buried and start
175. 5:39getting pushed to strangers. I know it
176. 5:40works because I've watched it happen
177. 5:42over and over. Let me show you. I didn't
178. 5:44learn from a course, I built my own
179. 5:46business from LinkedIn content. And then
180. 5:48I started doing it for clients. Between
181. 5:49me and them, that's over $4 million in
182. 5:51pipeline and more than 900 inbound demos
183. 5:54booked. All of it from LinkedIn content,
184. 5:56not one single cold message. Remember
185. 5:58the client from the start of the video?
186. 6:00This is her actual results. She's a
187. 6:01founder of a startup, 747,000
188. 6:04impressions in 28 days. That's up 206%.
189. 6:08But, here's the number that actually
190. 6:09mattered, and it should sound familiar
191. 6:10by now. 93% of that reach came from
192. 6:13outside of her network. Total strangers.
193. 6:15That's not luck. That's just the exact
194. 6:17thing LinkedIn's papers describe.
195. 6:19Human-specific content earns real
196. 6:21interaction, and the system pushes it to
197. 6:23people she's not connected to. I wrote
198. 6:24this post for the client last month, in
199. 6:26July, after all the algorithm changes.
200. 6:28On its own, it did 276,000 impressions,
201. 6:31and 96% of that came from out-of-network
202. 6:33connection. By that, I mean literally
203. 6:35people who don't follow her. Three moves
204. 6:37made that happen. You can copy all
205. 6:39three. Move one is the hook. "I was 23
206. 6:41when Deloitte offered me a job." That's
207. 6:42a single-line bomb. When you look at "I
208. 6:44was 23," right away, you unintentionally
209. 6:46started to think about what you were
210. 6:47doing at the age of 23. At that age,
211. 6:49most of us were broke, applying
212. 6:51everywhere, waiting to get an answer
213. 6:53from somebody that has all the power.
214. 6:55So, you are not just remembering an age,
215. 6:56you are literally remembering how it
216. 6:58felt. The job you wanted, the offer you
217. 7:00were waiting on, the one that came
218. 7:01through, or the one that got rescinded.
219. 7:03That memory shows up before you even
220. 7:04read the second line. Then, comes the
221. 7:06brand. Deloitte. Everybody knows that
222. 7:08name, and Deloitte itself has hundreds
223. 7:10of thousands of employees. Now, add
224. 7:11everyone else who ever interviewed
225. 7:13there, everyone who got rejected there,
226. 7:15and everyone who works at a firm just
227. 7:16like it. That is an enormous pool of
228. 7:18people who see that word and feel
229. 7:20something. At this point, we are playing
230. 7:22brand association on purpose, because a
231. 7:23name that big pulls its own audience
232. 7:25with it. So, by the end of sentence one,
233. 7:27a stranger is thinking about their own
234. 7:28memories at 23, and their own version of
235. 7:30the big firm that said yes or rather
236. 7:32said no. And they still have no idea
237. 7:34where this is going, because she opened
238. 7:36with a win. Deloitte offered her the
239. 7:38job, but something clearly went wrong,
240. 7:40and now you have to keep reading to find
241. 7:42out what. Move two, the three-beat drop.
242. 7:44Watch how fast this story turns.
243. 7:47Paragraph two is the dream. I remember
244. 7:48reading the offer letter and thinking
245. 7:50this was the moment that my career
246. 7:52finally cracked open. Paragraph three is
247. 7:55the reversal. Someone at Deloitte
248. 7:56realizes she never went to college and
249. 7:58they pull the offer after five
250. 8:00interviews. Paragraph four is the
251. 8:02landing. The offer that was going to
252. 8:03change my life disappeared in one single
253. 8:06email. The dream, the reversal, the
254. 8:08landing, three paragraphs, maybe 60
255. 8:11words, and do you feel all of it? Move
256. 8:13three is the payoff. Watch how she ties
257. 8:15the whole story together. Paragraph five
258. 8:17is the cost. My path to the C-suite
259. 8:19table was physically and emotionally
260. 8:20brutal. She does not skip the years she
261. 8:22spent proving herself to people who had
262. 8:24half her experience. She names them.
263. 8:27Paragraph six is the reliability factor.
264. 8:29Somewhere out there is a Latina in HR
265. 8:31who doesn't have a college degree. If
266. 8:33you're that exact person, she's calling
267. 8:35you out. They have no choice but to hit
268. 8:37follow and engage with the post.
269. 8:38Paragraph seven is the promise. None of
270. 8:41it is her ceiling. Her career is
271. 8:42possible. Her seat is coming. It's
272. 8:44short, it's certain, no hedging.
273. 8:46Paragraph eight is the address. Keep
274. 8:48going, amigos. I promise we make it. No
275. 8:51link, no pitch. She names exactly who
276. 8:53this is for and speaks straight to them.
277. 8:56Cost, reliability, promise, address.
278. 8:58That is why a stranger saves this and
279. 9:00sends it to somebody. And every line in
280. 9:02it came out of a real conversation with
281. 9:04her. My job was choosing what to cut,
282. 9:06not what to invent. Sure, AI can write
283. 9:08you a story about resilience, but it
284. 9:10can't write that story for you. And the
285. 9:12proof is in the pudding. 96% of Adam
286. 9:14Network connections saw this post, 249
287. 9:16new followers, all because it passed the
288. 9:18human test. So LinkedIn handed it to
289. 9:20strangers. So no, the algorithm did not
290. 9:22kill your reach because this post just
291. 9:24ran after it. So let's pull this all
292. 9:25together. LinkedIn didn't just tweak the
293. 9:27algorithm, they rebuilt it. And the
294. 9:29whole thing now asks one single question
295. 9:31for everything that you post. Did a real
296. 9:33human make this? And will real people
297. 9:35actually care? AI slop fails on every
298. 9:38level. It's sounds generated, it stays
299. 9:40trapped in your own network, and earns
300. 9:41nothing but a couple impressions here
301. 9:43and there. Human content does the
302. 9:45opposite. An original story, real proof,
303. 9:47your actual voice. It gets read, it gets
304. 9:49saved, it gets pushed to strangers. And
305. 9:51that's not theory, you just watched it
306. 9:52do 276,000 impressions. 96% to
307. 9:56strangers, too, on a post written
308. 9:57literally right after the algorithm
309. 9:59change. Now, if you want to learn how to
310. 10:00build a LinkedIn content system that
311. 10:02actually converts into pipeline, or you
312. 10:04just want to build your personal brand,
313. 10:05join the 4-week LinkedIn Accelerator
314. 10:07that we run monthly. If you'd rather we
315. 10:08just run your LinkedIn for you, that's
316. 10:10what we can help you handle. Both are
317. 10:11linked below, or send me a DM on
318. 10:13LinkedIn. And if you want to see more
319. 10:14content like this, don't forget to hit
320. 10:16the subscribe button, like the video,
321. 10:18and go ahead and comment. See you on the
322. 10:19next one, guys.