Which writing tactics actually work (analysis of 350,000 posts)

Across 350,000 LinkedIn posts, certain writing tactics consistently outperform others on likes. Here's the measured ranking — not guesswork.

In short

→ Social proof (citations, testimonials, third-party results): 36 median likes — the most engaging tactic measured. → Vulnerability (sharing a real mistake or failure): 33 likes. Authenticity beats opinion. → Quantified proof: 32 likes. A precise number stops the scroll and validates the argument. → Open-question CTA: boosts comments, but median likes drop to 25 — the lowest in the ranking. → Stacking: viral posts layer 4 to 6 tactics simultaneously. Single-tactic posts plateau.

The 6 tactics ranked by median likes

The analysis covers approximately 359,000 posts with complete metrics over 180 days, drawn from the 516,000-post LinkedIn corpus. Each tactic is detected automatically, then cross-referenced with 30-day likes.

| Tactic | Median likes | |---|---| | Social proof | 36 | | Vulnerability | 33 | | Quantified proof | 32 | | Polarization | 30 | | Comment-gate | 27 | | Open question | 25 |

Comment-gating scores 27 median likes — middle of the pack here. But it holds the record for median comments: 16, the highest of any tactic measured. It's not a virality tactic; it's a capture tactic.

Observational result, not causal: creators using these tactics may also have stronger hooks or more engaged audiences. Correlation, not causation. Keep sampling bias in mind.

Why social proof and vulnerability dominate

Both tactics share the same underlying mechanic: they anchor the post in an external, verifiable, or lived reality. Readers trust a third-party proof or a genuine admission more than a first-person opinion.

Social proof in practice: → "I tested this method with 3 clients — here are the results." → Third-party data, benchmarks, real case citations. → Unfiltered testimonials.

Vulnerability in practice: → "I got this wrong for 18 months before I understood why." → A concrete mistake — not generic false modesty.

Quantified proof (32 median) follows the same logic: a precise number is an instant form of social proof. In the top 1% of viral posts from the playbook, 61% include a precise number in the first 200 characters. That's not a coincidence.

The open-question CTA: likes vs. comments trade-off

This is the only genuinely counterintuitive finding. A closing open question produces the lowest median likes (25) but stimulates comments. The mechanic is simple: a direct question gives people a reason to reply. In exchange, readers who don't have an answer to give are less likely to like — because the post didn't ask them to.

In short: if you want likes, avoid the open-question CTA. If you want comments — for algorithmic velocity or DM capture — the question remains a useful lever, as long as you accept the trade-off.

For more detail: should you end your post with a question breaks down the numbers by CTA type and when to use each.

Polarization: strong on likes, not on comments

Polarization (30 median likes) beats comment-gating in raw likes but stays behind social proof and vulnerability. What changes with polarization is the ceiling: when backed by data, highly polarizing posts reach 2.75× more likes than the average.

The catch: polarization without evidence — a divisive opinion with no proof — doesn't get that multiplier. And the share of genuinely polarizing posts in the corpus is small. This is a high-risk, high-upside tactic, not a daily routine.

Stacking: the 4-to-6 tactic rule

No single tactic reliably creates a viral post. What sets the top 1% of posts apart is layering: a strong hook in the first 200 characters, quantified proof, a story or social proof anchor, long-form body, and a question or call-to-action at the close.

Single-tactic posts — just a number, just a question, just a story — stay in the average. It's the stacking of 4 to 6 elements that crosses the threshold.

That's exactly what the LinkPost virality score detects: 33 combined criteria, including the presence and density of these tactics in your text.

What to do

→ Ground your post in something real: third-party proof, a precise number, lived experience. Avoid generic opinions. → If you want likes, lead with social proof, vulnerability, or quantified proof. → If you want to maximize comments — for a lead magnet or capture mechanic — comment-gating is the most effective format. → Stack 4 to 5 tactics in the same post. Hook + quantified proof + vulnerability + long-form is already a strong profile. → Before you publish, run your post through the free post analyzer to see which tactics are detected and which are missing.

Observational study by LinkPost (2020 to April 2026, 62% French-language content, approximately 359,000 posts with complete metrics over 180 days). Results are correlations, not causal relationships. Full methodology and limitations in the complete playbook.