The best first line for a LinkedIn post, judged on 12,988 ranked posts

The best first line for a LinkedIn post is a short sentence with a person in it, describing something that already happened. Across 12,988 ranked posts, 19.6% of top-decile posts open in first person (I, my, we) against 10.1% of the bottom half, roughly double. That is the widest gap of any opening-word property we measured, and more importantly it is the only one that survives when we hold the subject matter constant: it points the same way in 16 of the 18 topics we can test. Questions, numbers and clever phrasings all fail that test. Who the sentence is about does not.

What counts as the first line, and how is it different from the hook?

Two different units get called the same thing, and conflating them is why so much opener advice contradicts itself.

The hookis everything LinkedIn shows before the “see more” cut, roughly 210 characters on the desktop feed and fewer on mobile. In our cohort its median length is 206 characters and 36 words, which is usually two or three sentences. The first line is sentence one: about a third of that budget.

This page is only about sentence one. Everything about the wider fragment, the fold mechanics, hard line breaks, how much of the preview a leading emoji costs, and how the opener changes by author role, is in how to start a LinkedIn post, which is the guide to the whole hook. If you have not decided what your post is about yet, read that one first. If you have a draft and the opening sentence is wrong, stay here.

The distinction matters because the two units get judged differently. The hook is judged by someone who has already paused. The first line is judged by someone still scrolling, and it is also the unit that gets pulled out and shown on its own in places that are not the feed at all.

What do 12,988 ranked posts say about the first line?

We filtered a public dataset of 34,012 LinkedIn influencer posts down to 12,988 English posts from 65 creators, scored each by engagement rate, and compared the top decile of 1,298 posts against the bottom half of 6,494. Three of the properties we measured describe the opening words specifically.

One row moves and two do not. First-person openers appear about twice as often in the winners. Opening with a question word moves 1.8 percentage points, which is close to nothing. Numbers move 0.7 points in the wrong direction, so the standard advice to lead with a statistic is, in this data, worth exactly zero.

The fourth row is the one people misread. Nearly half of top hooks put “you” somewhere in the visible fragment, and that gap is the largest on the chart at 14.8 points. But that is a property of the hook, not of sentence one. Directly addressing the reader in the very first words is genuinely rare in this cohort: in career growth posts, for instance, only 2.5% of the topic's own top decile (138 posts from 19 authors) open in second person, against 1.2% of its own bottom half (695 posts from 33 authors). The winning shape is not “you” first. It is a first line about a person, with the reader pulled in a sentence or two later.

Does the first-line rule survive inside a single topic?

This is the test most content studies skip, and it is the one that matters. A cohort-wide average can be an artefact of which subjects happened to do well. The stronger test holds subject matter constant: take one topic, split it into its own top decile and its own bottom half, and check whether the pattern still points the same way.

We ran that for every topic where both halves clear our floor of eight distinct authors. Here is the first-person opener rate in each topic's own winners against its own weakest posts.

Sixteen of eighteen. The two exceptions are narrow, and both are topics where a large share of posts are about an external event rather than the author: product news in technology, policy arguments in remote work.

Compare that with the same test applied to question marks, where the direction holds in 9 topics and reverses in 9, an exact coin flip. That contrast is the reason to trust the first-person finding and distrust most of the rest of the advice on this subject. One property behaves like a rule. The other behaves like an average. The full question-mark breakdown, including where questions genuinely do help, is in the question hooks analysis.

Note also the spread in the baseline column. A first-person opener is normal in startup posts (39.8% of the winners) and unusual in finance posts (12.9%). What travels is the direction of the gap, not the level. You do not need to write like a startup founder. You need to open more often in first person than the posts you are currently losing to.

Which first lines over-index in top posts?

Five shapes, all of which put a person in the opening words. Every example below is a real hook from the dataset with its real counts.

1. A moment with a timestamp on it

The first words locate the reader in time, and the sentence reports something that happened. No preamble, no explanation of why you are posting.

2. The improbable connection

Two things that do not obviously belong together, joined in the first sentence. The reader keeps going to find out how.

3. The direct address to a named group

Not “you” in the abstract, but a specific group called by name in the opening words. This is the rare case where second person in position one works, because the reader either is that group or is not, and both reactions are useful.

4. The reversal of your own stated position

Open with what you used to believe. The tension is built in, and it is hard to fake, which is part of why it reads as credible.

5. The visible nerve

The first line shows the writer's state rather than describing it. It works because it cannot be written by someone with nothing at stake.

None of these are templates. They are starting positions, and most strong first lines combine two: a moment that is also a reversal, or a nerve that is also an address. The longer-form version of these shapes, extended into full narratives, is covered in the storytelling guide.

Which first lines get skipped?

• The topic sentence.“Culture is built in the small moments.” True, general, and empty of people. Weak posts open with subjects. Strong ones open with someone doing something.
• The wind-up.Any sentence whose job is to announce that a post is beginning. “I have been thinking a lot lately about...” Delete it. Your post almost always starts at sentence two.
• The format announcement.“Here are 7 lessons from...” tells the reader the shape of the post before giving them a reason to care what is inside it.
• The bare statistic. Numbers appear in 32.1% of top hooks and 32.8% of weak ones. A figure with no story attached is not a hook, it is a fact.
• The credential. Your title is already on the card next to your face. Line one is the only sentence that has to earn the tap, so spending it on standing is expensive.
• The link or the hashtag block. Both push your actual first line out of the preview. The hashtag data argues for keeping tags minimal anyway, and the opening is the worst place for them.

Here is what the failure looks like in real data. This post opens with a label rather than a person, and the comment count tells the story.

How long should the first line of a LinkedIn post be?

Shorter than you think, but we should be straight about what the data can and cannot support here. Our source captures each post as a single truncated string with spacing normalised away, so we can measure the whole visible hook but not sentence boundaries inside it. Any specific word count for sentence one is craft advice, not a finding.

What the data does say is that total length is not the variable. The median hook runs 206 characters and 36 words in top-decile posts against 205 characters and 34 words in the bottom half. A one-character difference across 12,988 posts is not a signal. Long openers do not lose. Boring ones do.

The craft argument for a short first line is mechanical rather than statistical. The reader is deciding in under a second, a line they can take in at a single glance gets read, and a line that wraps to a third row usually does not. Around a dozen words is a reasonable ceiling. If yours runs longer, look for the word “that” or a subordinate clause and cut there. The full treatment of length, including where the body should end, is in how long a LinkedIn post should be.

Where else does your first line show up?

This is the part almost nobody accounts for, and it is a genuine argument for treating sentence one as its own unit rather than as the front of a paragraph. Your first line gets extracted and shown on its own in several places where the rest of the hook does not follow it.

The practical test that falls out of this: read your first sentence with everything after it deleted. If it is a fragment that only makes sense as a lead-in, it will look like a fragment in half the places it appears. If it stands alone as a statement someone might repeat, it works everywhere.

Does the first line work differently on a video or image post?

Yes, and the difference is large enough to change how you draft. On a text post, sentence one is the entire product. On a video or image post, it is a caption competing with a thumbnail that has already made its own impression. The measured behaviour of each format reflects that.

Two things stand out. First-person openers are most common on text-only posts and least common on link shares, which is close to a definition: a link share is usually about somebody else's work. And second-person mentions climb steadily as the post gets more visual, from 32.7% on link shares to 43.0% on video.

The practical reading is that a video caption should do a different job than a text post opener. The video is already showing the reader a face and a scene, so the caption does not need to build one. What it needs to do is tell the reader why this concerns them, which is why second person climbs. A caption that duplicates what the first three seconds of the video already say is wasted, and it is the single most common mistake in the format. The rest of the video picture, including why it over-indexes so heavily in the winners, is in the video posts breakdown, and the same logic applies to a still image with text baked into it.

How do you know your first line was the problem?

A post that underperforms gives you very little to work with, and people usually blame the wrong thing. The diagnostic is a ratio rather than a number.

• Impressions normal, engagement near zero. This is the first-line failure. A normal number of people were shown the post and almost none acted, which means the fragment they saw did not earn the tap. Rewriting the opener is the correct response.
• Impressions well below your own baseline. This is usually not the opener. Distribution was limited before the fragment had a chance to work, and the causes sit elsewhere: format, timing, posting pattern, or something on the account. Why a post gets no views works through the list.
• Reactions fine, comments absent. The opener worked and the ending did not. Comments are the scarcer signal in our data by a wide margin, at a top-decile median of 72 against 5 in the bottom half, so this is a real distinction rather than a rounding error.

The confusing part is that LinkedIn reports impressions and reach differently, so a comparison against last week's post can mislead you about which of these you are looking at. Check which figure you are reading before you conclude anything from it. Judge an opener over a run of ten posts rather than individually, because at ordinary follower counts a single engaged reader swings the numbers hard enough to make a good opener and a bad one look identical over a fortnight.

How do you fix a weak first line?

Most weak first lines are accurate summaries sitting where a hook belongs. The repair is almost always subtraction. The three pairs below are invented for illustration and are not from the dataset.

Cut to the event

Before:“I wanted to share some reflections on a hiring decision we made last year that taught me something important.”

After:“We hired the candidate who failed our take-home test.”

Put a person where the abstraction is

Before:“Psychological safety is the foundation of high-performing teams.”

After:“My best engineer told me she had been afraid to disagree with me for eight months.”

Move the surprise to the front

Before:“Over the last quarter we ran a series of pricing experiments, and the results were not what we expected.”

After:“We raised our price 40% and signups went up.”

In each pair the post underneath is identical. Only the entry point changed. Notice also what each rewrite has in common: a person or a first-person subject, a concrete verb, and something left unexplained.

The subtraction is worth dwelling on, because it is the part people resist. Every deleted clause was there for a reason. It explained context, softened a claim, or signalled that you know the subject is more complicated than one sentence allows. All of that is real, and none of it belongs in position one. The qualifications survive perfectly well in sentence four, where they are read by someone who has already decided to keep going. Put them in front and they are read by someone who has not, which is why hedged openers lose to blunt ones even when the hedged version is more accurate.

How do you write twenty first lines in twenty minutes?

Most people write the post and then hunt for an opener, which is backwards and produces summaries. The alternative is to collect first lines continuously and let them decide what you write.

1. Keep one running note. Every time you say something out loud at work that makes someone ask a follow-up question, write down the sentence you said, not the topic.
2. Capture the same day. The exact words go stale within about a week, and the specificity is the whole asset.
3. Write five candidates before you draft. First lines are cheap and post bodies are expensive, so generate options at the cheap end and let the winner decide what the post is about.
4. Run each candidate through the standalone test. Delete everything after it. Does it still mean something?
5. Check the subject of the sentence. If it is a concept rather than a person, rewrite it. That single check is the one property that held in 16 of our 18 topics.
6. Score it before you publish. The hook analyzer checks a draft opener against the patterns in this dataset.

Ten collected sentences is roughly two weeks of posts you do not have to invent from a blank page. If you want the raw material rather than the method, the career growth examples page lists real openers from the cohort with their counts attached.

What this analysis cannot tell you

These findings are correlational. Top posts open in first person about twice as often as weak ones, which is not evidence that changing your opener would have changed any particular post's outcome. The cohort skews toward established creators, all above 1,000 followers, across 65 accounts, so treat the baseline rates as directional if you are starting from zero. Engagement counts are lifetime totals at the time of collection and post ages vary, which adds noise. And every text finding describes the visible hook rather than the full post body, because the source data is truncated at the fold.

The topic table carries one more caution. Each row rests on a few dozen to a few hundred posts per half, so individual rows could move by several points. The reason we lean on it is the count rather than any single line: sixteen of eighteen pointing the same way is hard to get by accident, and it is exactly the test that most opener advice fails. The wider method, and every other property we measured, is in the full engagement study, with the pattern-level treatment in the hooks guide.