LinkedIn algorithm study: what actually worked across 130,000 posts
Flypost analysed 130,000 LinkedIn posts from 2,500 B2B creators and found that most common LinkedIn advice does not hold up. The useful finding is refreshingly simple: posts with images and first-person openers performed best. Many familiar tricks – hashtags, question hooks, perfect timing and ideal post length – either did nothing or made results worse.
Visit Flypost to read the whole fascinating article
Alternatively, read on for our ChatGPT quick summary version:
What worked best on LinkedIn
- Attach an image: +33% reactions compared with the same creator’s text posts.
- Open in first person: +20% reactions when posts started with I, my, we or our.
- Use a carousel: +13% reactions, but from a smaller sample.
- Use video: +11% reactions, again worth testing rather than treating as guaranteed.
The least fancy advice was the strongest. Add a relevant image and write like a person talking about something they did.
What is the best LinkedIn post format for engagement?
Start with image posts. They had the clearest lift in this study and worked for 70% of the creators tested. Carousels and video can help, but they were less reliable. The safer rule: use the format that makes the point clearer, not the one currently being hyped.
What did not work
- 1-2 hashtags: no change.
- 3-5 hashtags: -7% reactions.
- 6+ hashtags: -8% reactions.
- Opening with a question: -15% reactions.
- Ending with a question: no extra typical comments, and -10% reactions.
- Writing at the reader with you: -9% reactions.
- Adding a link: -12% reactions, though not a total disaster.
- Unpopular opinion hooks: no change.
- Opening with a statistic: no change.
Why the test method matters
Flypost compared each creator against their own posts. That is important. Comparing different people often gives misleading results because bigger creators, better writers or more experienced posters distort the picture.
Several popular claims fell apart once the same-person comparison was used, including long posts winning, Sunday being best, and frequent posters getting more engagement per post.
LinkedIn post length: no magic number
The famous 1,300-character rule was not supported. For text-only posts, length made no measurable difference. Across all formats, longer posts were only slightly ahead, and much of that seemed to be because longer posts were more likely to include images.
How long should a LinkedIn post be?
Write to the size of the idea. If the post needs 300 characters, stop there. If it needs 1,000, use them. A clear, useful post beats one stretched to hit a made-up character target.
LinkedIn posting time: probably less important than people think
Posting at the weekend instead of midweek changed engagement by about 2% for the same creator. That’s statistically insignificant.
- 60% of B2B posts go out on weekdays between 8am and 2pm.
- Only 4% of US and UK posts go out at weekends.
- Posting into a quieter slot did not reliably improve results.
The practical answer: post when you can do it consistently, then spend the saved overthinking time improving the post.
How often should you post on LinkedIn?
Posting more often reduced reactions per post, but lifted weekly reactions overall.
- Under 1.5 posts per week: 30 reactions per post, 18 per week.
- 1.5-4 posts per week: 20 reactions per post, 46 per week.
- 4+ posts per week: 13 reactions per post, 57 per week.
So if total reach matters, posting more can help. Just expect each individual post to look less impressive.
Should you post more often on LinkedIn?
Yes, if your goal is more total weekly engagement. No, if your only measure is reactions per post. Posting more spreads attention across more posts, so judge the week, not each single update.
What to do with links
Links were not banned, but they did carry a cost. A normal post with a link earned about 12% fewer reactions. A post that was mainly just a link share did worse, at around 16% down.
That makes links a trade-off. If the click matters, include the link. If the post is only there for reach, think twice.
Useful LinkedIn benchmarks
- The typical B2B post earned 34 reactions and 3-4 comments.
- Accounts with 5k-10k followers typically earned 27 reactions.
- Accounts with 25k-50k followers typically earned 43 reactions.
- Accounts with 100k-500k followers typically earned 143 reactions.
Comments are scarce. For smaller accounts, even a handful of real comments can be a decent result.
The practical takeaway
- Write from real experience.
- Use first-person openers.
- Add a relevant image where possible.
- Stop obsessing over hashtags.
- Do not force question hooks.
- Use links when clicks are the point.
- Post consistently, not perfectly.
The best lesson from the study is not an algorithm hack. It is that LinkedIn tends to reward posts that look and sound like they came from a real person with something useful to say.
