What 172 LinkedIn posts taught me about getting hired

I spent about seven months treating my LinkedIn feed like a lab. A couple months last year, then I got serious about it this year. I tracked 172 of my own posts. Impressions, saves, profile views, and comments for each one. I did it because I was doing two things at once. Looking for a senior marketing role, and building a product. Both needed the same thing. Getting found by the right people without paying for reach.

What I learned changed how I post. Most of it goes against the advice you see everywhere.

This is the whole system. It works whether you're a marketer, a nurse, an engineer, or anything else. But it only works if you make it yours, and I'll be straight with you about what that takes. There are two prompts at the end you can run on your own background today.

Start with the reframe, because this is where most job seekers go wrong. They treat LinkedIn like a place to repost their résumé and hope. That's not what gets noticed.

Why post at all when you're job searching

Apply cold and you're one of 300 résumés in a queue, sorted by software looking for a reason to cut you.

Get found because something you wrote made a recruiter stop scrolling, and you're a person with a point of view who happened to be exactly what they needed. They came to you. That changes the whole first call.

Posting does one job above all for a job seeker. It drives profile views from people who hire. So that's the goal. Not likes. Not going viral. Profile views from the right people, and saves from people who want to come back to what you said.

Hold onto that, because the next part goes against what the platform trains you to chase.

The two numbers that matter, and the one that lies

Across 172 posts, likes were inversely correlated with my best content. The posts that got the most likes were not the posts that got me profile views or saves. Likes feel good and tell you almost nothing.

Two metrics actually mattered:

• Profile views. When someone clicks from your post to your profile, they're evaluating you. Check your "who viewed your profile" list every week and look for Director-and-above titles in your field. That list is your warm pipeline.
• Saves. A save means "worth coming back to." Much stronger than a like, because the reader decided your content was useful enough to file away.

One catch from my data. The posts that drive profile views and the posts that drive saves are almost never the same posts. My top ten for saves and my top ten for profile views had zero overlap. So you can't write one kind of post and get both. You plan for each on purpose:

• Proof stories and sharp opinions get you profile views.
• Useful, tactical, how-to posts get you saves.

You need both. Know which one you're writing before you start.

The hook is the whole game

The biggest lever in my whole dataset was the first line.

• Strong hooks: median 2,646 impressions
• Weak hooks: median 829

That's a 3.2x gap, almost all of it from the first twelve words. Hooks over fifteen words were fatal. The reader decides whether to open your post in about a second, off one line.

What worked, ranked by how it performed for me:

• A hook that names a specific identity. It makes one kind of person think "that's me, that's my exact situation." My top performer.
• A hook that names dysfunction or friction. A close second, and the best hook for any opinion post. When I led with a specific broken thing I'd seen, those posts hit 4.7x higher than any other opening.

What failed every time:

• Opening with a question
• Opening with your title or years of experience
• Opening with a flat setup line that just announces the topic

The rules, plain:

• Lead with tension, friction, or a specific identity
• Never open with a question
• Never open with your credentials
• Keep it under twelve words
• Start with "I" or "you"
• Put your single best line first. If the good part is in paragraph three, move it up. I lost reach for months burying it.

How to build the rest of the post

Once the hook earns the click, the shape carries it:

1. Hook. A confession, a claim, or a named problem.
2. Context. One sentence on why you're the one saying this.
3. The substance. The story, framework, or diagnosis, specific enough to use.
4. The personal part. What you tried and what happened. This separates you from a content account.
5. A specific question. One that invites the exact person you want to hear from.

Hard limits from the data:

• Aim for 250 to 350 words. That range won on reach and saves.
• Under 150 words isn't a post, it's a comment. Over 450, cut.
• Default to story, or story plus a little structure. Plain lists of insights underperformed.

The one rule I'd put above the rest: demonstrate, don't describe. Show the real number, the actual moment, the exact line you changed. "Here's the email line that tanked our open rate and what I swapped it for" beats "email tone matters." Every time.

And stop closing with "Thoughts?" In my data it pulled a median of half a comment. Write the question for the person you want instead:

• "Which part of the job search is breaking you right now?" pulls fellow job seekers.
• "What's the engagement problem you're fighting on your team?" pulls the manager who has it.

The question's job is to choose who shows up in your comments, not to rack up a number.

Know exactly who each post is for

Before you write, answer two things. Who is this for, and what do I want them to do. "Everyone" is not an answer. A post for everyone lands with no one.

The answer shifts by post type. A proof story is for a hiring manager in your field, and you want them on your profile. A how-to is for a peer who'll save it and remember your name. A domain opinion is for the person who lives that broken thing every day and finally sees someone name it. Write each post to one of those people, not to the room.

This is why the closing question carries so much weight. It's where you signal who should answer. Pick the person first. Then write the whole post to them.

The four kinds of posts, and the mix

Every post does one of four jobs. Know which before you write.

• Proof stories. A specific thing you owned. The problem, the fix, the outcome, ideally with a number. Lead with the outcome, never the backstory. This is your profile-view engine. Most job seekers run none of these because it feels like bragging. Run one a week anyway.
• Domain point of view. A sharp, specific take on something broken in your field, grounded in what you saw firsthand. Not a news recap. These build authority. They need a dysfunction hook or they die.
• Useful how-to. A framework, checklist, or system from your own field that someone can use in five minutes. This is your reach engine. It gets saved and shared, which makes it travel, which grows the audience your proof stories and opinions land in front of later. Keep it tied to your real expertise, not generic job-search hacks, or you'll build an audience that can't hire you.
• Personal and identity. A post with real professional stakes. A pivot, a hard moment, a choice that shows your character. Not lifestyle. These drive the most comments and they work through advocacy. They make people think "I should introduce them to someone." That referral path is one of the most valuable things on the platform.

A good week is four to five posts, one a day, and the mix matters more than the count. Some posts are built for reach. Some are built to get the right person onto your profile. You need both running, because the first kind feeds the second. Only have three good posts in you this week? Run three.

Cadence and mechanics, where most effort gets wasted

Short section. Simple rules. Breaking them costs more than people think.

• One post a day. Hard cap. My second post on any day lost 30 to 80 percent of the first one's reach. Every time. And more posting didn't mean more results. My five-post weeks beat my ten-post weeks on engagement (1.10% vs 0.59%) and follower growth (+177 vs +94). Volume divides your results. It doesn't multiply them.
• Post in the morning. Mine did best around 8:45 to 10am. Test your own window, but lean early.
• Be there for the first 60 to 90 minutes. Early comments are the gas pedal. They tell the platform to keep showing your post. Reply to everyone in that window. If you can't sit with the post for an hour, don't post that day.
• No weekends. And don't post right before an interview. You need that hour on the interview, not refreshing comments.

One more I learned the hard way. When a post breaks out and reaches way more people than usual, the next thing you post is your most important post of the week, because all those new people are about to meet you through it. Make it a proof story or a strong opinion. Don't waste it.

Stop doing these

These dragged down everything that came after them in my data:

• Reposting other people's content, whether it's a plain repost or a "Love this!" with no take of your own. Nobody comments on those, and comments are the whole trigger you're trying to earn.
• Pure excitement with no tension, the "so thrilled to share" post
• Corporate announcement voice. The press-release tone got me zero engagement, no exceptions.
• Opening expertise posts with a question or your credentials

Don't sound like a robot

You can write a perfect post and still sound like AI wrote it. People feel it even when they can't name it, and it costs you trust. Run everything through this before you publish:

• No em dashes. Ever. One of the clearest tells right now. Use a period or a comma.
• Use contractions. "I've," "don't," "I'd." Writing "I have not" reads stiff.
• Vary sentence length. AI writes long sentences of the same size, over and over. Real writing punches short, runs long, lands short.
• Be specific. Name the company, the number, the moment. Generic claims are what bots make.
• Cut filler. "Dive into." "Leverage." "Passionate about." "At the end of the day." If you've read it a hundred times, it's noise.
• Own gaps without apologizing. "I haven't done that part, but here's what I have done" reads human. Hedging every sentence reads like a machine covering itself.

The test: would someone who knows you believe you wrote it? If not, it's not ready.

The part you can't copy

You can follow every rule in here and still get nothing, because the rules are the scaffold. They're not the substance. The substance is you.

My posts that worked were the ones where I said something I was a little nervous to put my name on. The campaign I ran that missed, and what it taught me. The opinion I knew some people in my field would push back on. Those did the work. The safe, polished, could-have-been-anyone posts did not.

People can feel the difference between someone reporting a framework and someone who lived it. The framework is borrowable. The scar is not. So take this whole system and run your own material through it. Your real wins. Your real opinions. The thing you actually believe that most people in your field won't say out loud. A little vulnerability is the whole signal.

Copy my structure. Don't copy my stories. Tell yours.

Two prompts to build your own version

Paste these into any AI. The first helps you find your material. The second takes the rules that worked for me and helps you turn them into your own playbook.

Prompt 1. The interview, to find your content.

Interview me to figure out what I should post on LinkedIn while I job search. Ask one question at a time and wait for my answer before the next. Start broad (my field, the role I want, who I want to notice me), then go specific: a project I owned and the outcome, a number I'm proud of, something broken in my field I've watched up close, a moment my career shifted, an opinion others in my field would push back on. After eight to ten questions, stop and give me three proof-story ideas, two domain-opinion ideas, one useful how-to idea, and one personal-identity idea. For each, write a one-line hook under twelve words that leads with tension or identity and never opens with a question.

Prompt 2. Build your own playbook.

Here are the rules that worked for someone who tracked 172 of their own LinkedIn posts as a job seeker. Help me adapt them into my own playbook, then check my drafts against it.
Goal: profile views from people who hire, and saves. Ignore likes. They're inversely correlated with good content.
Audience: before each post, make me name exactly who it's for and what I want them to do. That changes by post type. "Everyone" is not allowed.
Hooks: under twelve words, lead with tension, friction, or a specific identity, never a question, never my credentials. Put my single best line first.
Structure: 250 to 350 words. Open with the hook, one line of context, the substance, what I personally tried and what happened, then a specific question that invites the exact person I want. Demonstrate with real numbers and moments. Don't describe concepts.
Mix: rotate four post types. Proof stories (a problem I owned and the outcome, lead with the outcome). Domain point of view (a specific broken thing in my field, dysfunction hook). Useful how-to (a framework from my expertise, drives saves). Personal identity (a real professional moment, not lifestyle).
Cadence: one post a day, four to five a week, mornings, no weekends. Don't close with "Thoughts?"
Voice: no em dashes ever, use contractions, vary sentence length, be specific, cut filler like "dive into," "leverage," and "passionate about." It should read like a human wrote it.
When I paste a draft, tell me: who is this for and is that clear, is my best line first, is the hook under twelve words and not a question, is it 250 to 350 words, does it demonstrate or just describe, and are there any em dashes or AI phrases to cut.

Where the flywheel points

Not every post is chasing the same thing. Some are there to be useful and travel wide. Some are there to get one hiring manager onto your profile. Some are there to make a former colleague think of you when a role opens. Different jobs, on purpose.

But they compound. A useful post grows your audience. A bigger audience means your next proof story reaches more of the right people. More of the right people reading your posts means more of them clicking through to your profile. That's the flywheel. It takes a few weeks to start spinning, and then it carries you.

The flywheel has one leak, and most people never find it. Everything you post points traffic toward one place: your profile. And a profile visit only becomes a conversation if two things are true when someone lands.

• A recruiter could find you in search in the first place. That depends on your keywords, your headline, the way your target role is written.
• Your profile reads like a person who solves their problem, not a keyword-stuffed résumé that sounds like everyone else's.

Most profiles fail at least one, and you never find out, because nobody tells you why they didn't reach out. You just stay invisible while the flywheel leaks.

So I built JobSearchDx for it. It runs your profile through the same diagnostic I use, shows you where you're invisible to recruiter search, and tells you what to rewrite so it reads human. It costs about the price of a coffee and a cookie. I put an early version out with almost no promotion and it got around 1,300 views, which told me how common the problem is.

Build the flywheel. Then make sure it isn't leaking. You can check your profile here.

And start posting like you have something to say. You do. The people you want to work for are scrolling right now.