LinkedIn Engagement by Post Type
Post type separates LinkedIn accounts more sharply than almost anything else in the data. Posts carrying more than one media item run a median of 99.8 reactions and comments, posts with a single media item run 43, and text-only posts run 24. Those figures are computed from 66,085 posts across 449 distinct accounts, with each account's own median taken first. The gap between the top and bottom type is roughly 4 times, which is larger than the gap between any two industries in published reporting.
Last updated: September 2026 · Next update: December 2026 · Methodology
The three measured post types, median account
More than one media item
99.8
90 accounts, 2,275 posts
One media item
43
392 accounts, 29,589 posts
Text only
24
411 accounts, 30,778 posts
Engagement per post by post type
| Post type | Bottom 25% | Median |
|---|---|---|
| More than one media item | 50 | 99.8 |
| Repost of another post | 40 | 86 |
| One media item | 15.5 | 43 |
| Text only | 9 | 24 |
Computed on 2026-09-09 with a within-account query that mirrors benchmark_within_sender(): each account's own median per type first, then percentiles across accounts. An account must contribute at least 10 posts of a type to enter that row, and a row needs at least 20 accounts to be published. Polls fell below both floors and are not shown. This is a media-count split, so it does not identify carousels or documents specifically.
What the media split looks like without counting items
| Cut | Bottom 25% | Median |
|---|---|---|
| Any media attached | 22 | 57 |
| No media attached | 9.4 | 27.8 |
| All posts | 12 | 34 |
| All posts excluding reposts | 11 | 34 |
The has-media split is covered in depth on the images page, which is why it appears here only as the frame around the post-type table. The interesting result is that counting media items separates accounts further than a simple yes or no on media does.
How to read the post-type table honestly
Reposts are a counting artefact, not a content lesson
A repost inherits the counters of the post it shares, so the 86 median reflects the original author's audience rather than the reposter's. Rewriting the idea and publishing it fresh performs better than resharing it, and the reposts row is published here only so the number is visible rather than hidden.
Multi-media posts come from a narrow set of accounts
Only 90 accounts contribute enough multi-media posts to enter that row, against 411 for text-only. Accounts that build multi-image posts tend to be the ones already investing in production, so part of the 99.8 median belongs to the account and not to the format.
The media flag does not name the format
Creator's corpus records how many media items a post carries, not whether they are photos, a document or a video. Any source giving you a precise carousel against image multiplier should be asked what field it read to tell them apart.
Comments move more than reactions across types
Median comments run 3 on text-only, 7 on a single media item and 12 on multi-media. That is a 4 times spread on comments against a roughly 4 times spread on total engagement, so the format lift is genuinely conversational rather than a wave of extra one-tap reactions.
How to change your post-type mix without dropping quality
- 1
Measure your current mix first
Count how many of your last 30 posts carried no media, one item and more than one. Most accounts discover they are far more text-heavy than they assumed. In the corpus, text-only and single-media posts are close to an even split at 30,778 against 29,589.
- 2
Move the weakest text posts, not the best ones
Take the five posts from your last 30 that landed below your own median and rebuild those as visual posts. Your strongest text posts are working, and adding an image to them is a smaller gain than fixing the ones that failed.
- 3
Add a second media item to one post a week
The measured step from one item to more than one is 43 to 99.8. Test it on one post a week for six weeks before drawing a conclusion, because a 90-account row is thinner evidence than a 411-account row.
- 4
Write the post before you make the visual
The comments column moves with type because multi-media posts usually carry a structured argument. A visual attached to a thin post moves reactions and leaves comments at 3.
- 5
Judge on 20 posts per type
Each account in these rows contributes at least 10 posts of a type. Give yourself at least that many before deciding a format does not work for your audience.
Frequently asked questions
Does the number of images in a LinkedIn post change engagement?
Yes, and the step is large. Posts carrying more than one media item run a median of 99.8 reactions and comments, against 43 for a single item and 24 for text only, computed across 66,085 posts and 449 accounts. Only 90 accounts contribute enough multi-media posts to enter that row, so treat it as a strong signal from a narrow sample rather than a guarantee.
How much does adding a second image change LinkedIn engagement?
In the measured data the median moves from 43 to 99.8 when a post carries more than one media item instead of one, roughly 2.3 times. Median comments move from 7 to 12 over the same step. The sample behind the multi-media row is 2,275 posts across 90 accounts, so test it on your own account across at least 20 posts before treating it as settled.
Do reposts get more engagement than original posts?
They appear to, and the appearance is a counting artefact. Reposts in the corpus show a median of 86 reactions and comments across 103 accounts, but a repost inherits the counters of the post it shares, so the figure reflects the original author's audience. Excluding reposts entirely leaves the corpus median unchanged at 34 across 63,371 posts, which is the cleaner number to plan against.
How many post types should I rotate through?
Two or three is enough. The measured spread comes from media count rather than from novelty, and accounts that hold a steady format mix accumulate a recognisable pattern in the feed. A workable weekly shape is two single-media posts, one text-only post for short personal notes, and one multi-media post when the argument needs steps.
Why does the same post type perform differently for different accounts?
Because the account carries more of the result than the format does. The bottom quartile of multi-media posters sits at 50 while the top quartile of text-only posters sits at 89.5, so a strong text account beats a weak visual account. Format changes your ceiling. Consistency and audience fit decide where inside that ceiling you land.
Does post type matter more than posting volume on LinkedIn?
They are close. Moving from text-only to multi-media is a step from 24 to 99.8 in the measured medians. Moving from the corpus baseline to accounts contributing 100 or more posts is a step from 34 to 54. Format has the larger single jump, volume is easier to sustain, and the accounts in the top quartile do both.
Can Creator tell a carousel apart from an image post?
No. The corpus records how many media items a post carries, not whether they are photos, a document or a video. The post-type table on this page is a media-count split and is labelled that way. Any source quoting a precise carousel against image multiplier should be asked which field distinguished them.
Do text-only LinkedIn posts get any comments?
Yes, at a median of 3 per post against 7 for single-media posts and 12 for multi-media, measured across 411, 392 and 90 accounts respectively. Text-only posts still work for short personal notes where a visual would look staged. As a default setting for every post, they start at roughly a quarter of the multi-media ceiling.
Methodology
Post-type figures come from a within-account query over Creator's LinkedInPost table, run on 2026-09-09, that mirrors the benchmark_within_sender() method: each account's own median per post type first, then the 25th, 50th and 75th percentile across accounts.
An account must contribute at least 10 posts of a type before entering that type's row, and a row needs at least 20 accounts before it is published. Poll posts fell below both floors and are withheld rather than published with a caveat.
Sample: 66,085 posts across 449 distinct accounts. That account count is higher than the 433 used on the whole-corpus pages because the per-type floor is 10 posts of a type rather than 30 posts overall, which lets in accounts that post rarely but consistently in one format.
Post type is derived from three stored fields: whether the post is a repost, and how many items sit in the media array. It records media count, not media format, so the table cannot separate documents, photos and video.
The has-media rows come from benchmark_within_sender() with the media filter set, giving 33,242 posts across 333 accounts with media and 30,871 posts across 324 accounts without.
Posts are B2B founder, consultant, agency and operator profiles. Engagement is read at scrape time, so a post can keep accruing after we count it, which biases every figure slightly low rather than high.
All rows on this page are computed from the Creator LinkedInPost dataset on 2026-09-09, using benchmark_within_sender() for the has-media rows and a matching within-account query for the post-type rows. No compiled range appears on this page.
Understand the terms
Carousel Post
A multi-page document post swiped in the feed.
Text Post
A post published with no attached media.
Native Video
Video uploaded to LinkedIn rather than linked out.
Poll Post
A post asking readers to vote between options.
Engagement Rate
Reactions plus comments plus shares, divided by impressions.
Dwell Time
How long a reader stays with a post before scrolling.
LinkedIn Content Playbook
The weekly system behind the accounts in the top quartile of this dataset.
More LinkedIn benchmarks
Do Images Increase LinkedIn Engagement? →
The media split in depth: 57 against 27.8.
LinkedIn Impressions by Post Format →
The same question measured on reach instead of engagement.
Comments or Reactions: Which Should You Track? →
Comment share of engagement, measured across 433 accounts.
LinkedIn Carousel Benchmarks →
Format-level figures for document posts.
See how your content compares
Creator tracks your performance against these benchmarks automatically and tells you exactly what to fix.
Apply to work with us