The structure of a LinkedIn post that works (hook, body, CTA)

A LinkedIn post that works always follows the same framework. We measured it across 438,413 posts: hook first, scannable body, CTA matched to your goal — in that order.

In short

→ The first 200 characters decide whether LinkedIn amplifies your post or buries it. → Opening with a precise number gives +17% median likes. Opening with a question loses -28%. → A TOFU CTA (value, education) generates +44% median likes vs. a direct-sell post. → One idea per post, short lines, generous white space — mobile-first by design. → Your CTA sets the objective: engagement, leads, or reach. Never try to do all three at once.

Why structure matters more than substance

LinkedIn doesn't read your post — it measures how the first readers react. If scrolling stops and people read, it amplifies. If they keep scrolling, your content gets buried before anyone had a chance to judge it.

Your structure does this filtering before your ideas even get evaluated. A brilliant post with a weak hook never gets read. A decent post with solid structure clears the filters.

The good news: structure is learnable, testable, and iterable.

The hook: the first 200 characters, nothing else

The hook is the part visible before "…see more." On mobile, that's 2 to 3 lines. This is where you lose or keep your reader.

Here's what the data shows on opening type:

| Opening | Median likes | |---------|-------------| | Starts with a number | 28 | | Neutral (no marker) | 24 | | Starts with a question | 18 |

Opening with a precise number gives +17% median likes. Opening with a question loses -28%. A question in the first line reads as hollow before you've proven anything.

In our study of 516,144 posts, the top 1% of viral posts share two traits: a strong hook in the first 200 characters, and a quantified proof point (61% of them include one).

Two hook formats that consistently work: → Number + contrast: "82% of LinkedIn posts gain zero followers. Here's why." → Polarizing claim with an anchor: "I analyzed 3 years of posts. Tuesday 9am does nothing."

To go deeper, the article on LinkedIn hooks breaks down 12 hook types with their measured performance.

The body: one idea, white space, scroll

The body of a LinkedIn post isn't a compressed blog article. It's one idea developed along a single axis, with as little reading friction as possible.

Practical rules: → One idea per post. If you want to say two things, write two posts. → Short lines: 1–2 sentences max per line — no walls of text. → White space between blocks: it helps scanning on mobile and gives the vertical read a rhythm. → Descending logic: most important first, least important last. The reader can stop at any point.

The writing tactics measured across 350,000 posts show that posts combining multiple signals — real anchoring, social proof, vulnerability — have significantly higher like medians than those applying just one. Signal stacking works at the body level as much as at the hook level.

The CTA: TOFU wins, BOFU loses

The end of your post sets the objective. And the algorithm judges this part too.

TOFU posts (education, value, inspiration) have a median of 36 likes vs. 25 for BOFU posts (direct selling, service promotion). That's +44% in favor of content that doesn't sell upfront.

This isn't a call to never sell. It's a calibration: sell in 1 out of every 5 posts at most, and always after delivering value first.

Three CTA types depending on your goal:

→ Engagement: an open question on the topic. Place it at the end of your post — not on the first line (see the data on question openings). → Leads: "Comment [keyword] and I'll send you [resource] in a DM." The comment-gate mechanic has the highest median comment count of all measured tactics. LinkMagnet automates lead magnet delivery in DM the moment someone comments the keyword. → Discovery: a link to deeper content (article, newsletter). Watch out — external links reduce impressions, so use sparingly.

What to do, concretely

1. Write your hook last: once the body is done, you know exactly what to promise upfront.
2. Start with a precise number: concrete, dated if possible, uncomfortable if you want reach.
3. Read your post on mobile before publishing: if you have to scroll four times to get through it, your readers won't.
4. Calibrate your CTA to one single objective per post — not three.
5. Check the structure in the free LinkPost analyzer before publishing: it detects weak signals (hook too long, missing proof point, single-tactic structure) and gives you a score across 33 criteria.

Structure doesn't replace ideas. But it guarantees your ideas have a chance of being read.

Observational study by LinkPost (2020 to April 2026, 62% French-language content, 438,413 posts). Findings are correlations, not causal relationships. Sampling bias and methodology detailed in the playbook.