Live data100,000+ LinkedIn posts, updated monthly

The #1 LinkedIn
post format in 2026

We analyzed 100,000+ LinkedIn posts. The data on what actually wins by reach, engagement, and comments.

first, a quick note

MyFeedIn tracks every post and every comment for thousands of LinkedIn creators. Over time, that adds up to a pretty wild dataset. We pulled 100,000 posts and looked at which formats actually win on impressions, likes, and comments. Here's what we found.

the data

Averages across all 100K+ posts

Format leaderboard

FormatImpressionsLikesComments
Video
1,76223219
Carousel
1,1696316
Image
5468010
Text
281244

Impressions by format

Last 12 months

Text
Image
Video
Carousel

Head-to-head comparison

How each format stacks up

ok, but here's the thing

Real talk

Yes, video wins on the chart.
So what?

Charts don't post on LinkedIn. You do. And the format that wins on a chart isn't worth much if you can't actually keep doing it every week.

The best format for you is the one you can repeat without burning out.

If you suck at design, don't do carousels. A bad carousel hurts your brand more than no carousel does. If you cringe on camera, don't do video, you'll quit in 6 weeks and reach zero. If writing feels easy, ride that. Text-only posts get the most comments anyway.

Pick the format that fits your strengths and your week. Then ship it consistently. That beats chasing the algorithm every single time.

Repeatable

Can you do it every week for 6 months?

Bearable

Does making it feel light, not draining?

Profitable

Are you seeing some signal after 4-6 weeks?

so which one is right for you?

Pick the one you can actually keep doing

Answer the question for each. The honest answer.

Text-only

Comfortable writing in your own voice?

Yes

Lean in. Text posts get the most comments, and that's where real connections happen.

No

It's the easiest format to start with. You can always rewrite. Practice for a month.

Image

Have something visual to show (work, results, screenshots)?

Yes

Use it. Image posts are the safest balance of reach and engagement.

No

Don't force a stock photo. The image has to add something or it drags the post down.

Video

Comfortable on camera, OK with editing?

Yes

Go video. It pulls the most reach right now and it's harder for others to copy you.

No

Skip it. A bad video performs worse than a good text post, and you'll burn out trying.

Carousel

Decent at design (or have a tool for it)?

Yes

Do carousels. They get saved and shared, which compounds your reach over weeks.

No

Don't bother. A bad-looking carousel hurts your brand more than no carousel does.

honestly, that's it. pick one and post for 90 days.

See this data for your own posts.

MyFeedIn breaks down impressions, likes, and comments by format for every post you publish. So you stop guessing what works and start posting what does.

Quick answers

The questions we get the most about all this.

Honestly? The one you can keep doing every week without hating it. Video pulls the most reach in the data, but if you're not comfortable on camera, you'll quit in 6 weeks and reach zero. Pick the format that fits your strengths and your week.

Numbers come from real LinkedIn posts tracked by MyFeedIn users. We measure average impressions, likes, and comments across thousands of posts, grouped by format (text, image, video, document) and aggregated monthly. The dataset refreshes daily.

LinkedIn's algorithm reweights formats periodically. Video had a strong 2023, document posts surged in 2024, and image posts have stayed steady throughout. The trend lines on this page reflect those shifts in real time.

No, but pick a primary one. Most successful creators have one format they ship every week (their main thing), and they sprinkle in a second format when it makes sense. Trying to do all four is how people burn out.

Yes. MyFeedIn tracks the same metrics for every post you publish, broken down by format. You see your own impressions, likes, and comments by format type, and which posts beat your own average.

Ready to track your own LinkedIn data?