Do LinkedIn Posts With Images Get More Engagement?
Yes, and the gap is 2.1×. LinkedIn posts carrying an image, document or video earn a median 57 reactions and comments per post, against 27.8 for posts with no media. The media figure is computed from 33,242 posts across 333 accounts and the text-only figure from 30,871 posts across 324 accounts, and the gap holds in every cut we tried.
Last updated: September 2026 · Next update: December 2026 · Methodology
Median engagement per post, with and without media
Posts with media
57
33,242 posts across 333 accounts
Posts without media
27.8
30,871 posts across 324 accounts
The gap
2.1×
Median account, reactions plus comments
Media against text-only, by metric
| Metric | With media | Without media |
|---|---|---|
| Reactions plus comments | 57 | 27.8 |
| Reactions (likes) only | 41.5 | 22 |
| Comments only | 10.5 | 4 |
All six figures are median accounts, computed on 2026-09-09. Media lifts comments more than it lifts reactions, 2.6× against 1.9×. An image stops the scroll and gives people something specific to respond to.
The gap holds in every cut we ran
| Cut | With media | Without media |
|---|---|---|
| All posts | 57 | 27.8 |
| Published 2025 onward | 74 | 37 |
| English-language posts | 75.5 | 35 |
| Excluding reposts | 56 | 26 |
Four independent calls to benchmark_within_sender(), each re-deriving the whole distribution over the posts that match. The gap moves between 2.0× and 2.2× across all of them, which is why we publish it as a finding rather than a coincidence.
Which media format to reach for (compiled ranges)
| Format | Effort |
|---|---|
| Document or carousel | High |
| Single data image | Medium |
| Real photo of you | Low |
| Native video | High |
| Screenshot | Very low |
Format guidance is compiled from Creator's own content operations, not computed. Our dataset records whether media was attached, not which kind, so we do not publish a computed ranking of formats and will not invent one.
Why media moves the number
It buys dwell time, and dwell time buys distribution
A reader who stops to look at an image spends longer on the post. Longer time on a post is one of the signals that decides whether LinkedIn shows it to a second wave of people. The reactions follow the reach rather than causing it.
It gives people something concrete to answer
The comment gap is the largest of the three, at 2.6×. A chart, a screenshot or a slide gives readers a specific thing to agree or disagree with. A block of text gives them a sentiment to agree with, and agreement produces a reaction rather than a comment.
It survives the truncation point
A text post is cut off after a few lines on mobile and the reader has to tap to see more. An image sits below the fold-line and stays visible either way, so it keeps working on people who never expand the post.
It does not rescue a weak idea
The bottom quartile of media posts sits at 22 reactions and comments, well below the median text-only post at 27.8. Adding an image to a post nobody wanted to read moves it from bad to slightly less bad.
How to close the media gap on your own account
- 1
Count your own split first
Take your last 30 posts and mark each one as media or text. If fewer than half carry media, the 2.1× gap is the largest single lever available to you.
- 2
Attach media to the posts you already know work
Start with your best-performing text posts from the last quarter. Rebuild each one with a single image that carries the main claim. You are testing the format change, not a new idea.
- 3
Make the image readable at thumbnail size
One claim, large type, high contrast. If you have to zoom to read it on a phone it will be scrolled past. Save the detail for the post body.
- 4
Use documents when the idea has steps
A framework, a checklist or a sequence belongs in a document. A single number or a single claim belongs in one image. Choosing the wrong container is the most common reason a well-made visual underperforms.
- 5
Re-measure after 20 posts, split by format
Compare your media median against 57 and your text-only median against 27.8. Judging both against one blended average hides which half of your content is doing the work.
Frequently asked questions
Do LinkedIn posts with images get more engagement?
Yes. Posts with media earn a median 57 reactions and comments per post against 27.8 for posts with no media, a 2.1× gap. The media figure is computed from 33,242 posts across 333 accounts and the text-only figure from 30,871 posts across 324 accounts, with each account's own median taken first.
How much more engagement do LinkedIn images get?
About twice as much. The gap is 2.1× on reactions plus comments, 1.9× on reactions alone and 2.6× on comments alone. It stays between 2.0× and 2.2× when the data is restricted to 2025 onward, to English-language posts, or to original posts with reposts removed.
Do images increase comments on LinkedIn?
More than they increase reactions. The median account gets 10.5 comments on posts with media and 4 on posts without, a 2.6× gap, while reactions move only 1.9× from 22 to 41.5. A visual gives readers a specific claim to respond to instead of a general sentiment to agree with.
Is it better to post text-only on LinkedIn?
Not on this evidence. Text-only posts run a median of 27.8 reactions and comments against 57 with media, across 324 and 333 accounts respectively. Text-only still works for short personal notes where an image would look staged, but as a default setting it halves the ceiling.
What kind of image works best on LinkedIn?
Creator's dataset records whether media was attached, not which type, so there is no computed ranking of formats here and we will not invent one. From our own content operations: one claim in large readable type, a real photo rather than stock, and documents rather than single images whenever the idea has steps.
Do carousels get more engagement than single images?
That comparison is not computable from this dataset, because the media flag does not distinguish document posts from image posts. Creator publishes a separate carousel benchmark page for format-level figures. Treat any source claiming a precise carousel-against-image multiplier with the same suspicion, and ask what sample it came from.
Will adding an image fix a post that flopped?
No. The bottom quartile of posts with media sits at 22 reactions and comments, which is below the 27.8 median for text-only posts. Media raises the ceiling on an idea people already want. It does not create demand for one they do not.
Does video beat images on LinkedIn?
Creator's computed figures group all attached media together, so video is inside the 57 median rather than separated from it. Anyone quoting a precise video-against-image multiplier is working from a different dataset, and the honest answer from ours is that any media beats none by 2.1×.
How many of my posts should carry an image?
In this dataset the split is close to even, 33,242 posts with media and 30,871 without. Given the 2.1× gap, a working default is media on every post that has something to show, and text-only reserved for short personal notes where a visual would be filler.
Methodology
This page counts reactions and comments per post. It does not measure engagement rate, which divides that count by impressions. The two move differently: a post can collect more reactions while converting a smaller share of the people who saw it. Our /stats/linkedin-carousel-benchmarks page reports rates by format and shows a different ordering for that reason.
Media here means any attached image, document, carousel or video, because that is the distinction our data records. It is not a single-image bucket, and the media group skews toward documents and video. Read the 2.1x as media against plain text, not as image against text.
Both figures come from the same live RPC call, benchmark_within_sender('reactions_plus_comments', p_has_media => true) and the same call with false, run against Creator's production database on 2026-09-09. Media means the post carries a non-empty media array: image, document or video.
Sample: 33,242 posts across 333 distinct accounts with media, and 30,871 posts across 324 distinct accounts without. The two account sets overlap heavily, because most accounts publish both kinds, which is what makes the comparison meaningful rather than a comparison of two different populations.
Each account contributes its own median before percentiles are taken across accounts. A single account posting 400 image posts cannot move the media figure, and a single viral post cannot move either.
The gap was re-run on three further cuts to test it: posts from 2025 onward (74 against 37), English-language posts (75.5 against 35), and original posts with reposts excluded (56 against 26). It landed between 2.0× and 2.2× every time.
The media format table is compiled guidance from Creator's own content work, not a computed ranking. The dataset stores whether media was attached, not which kind, so no format-level figure is published here.
Every number in the headline, the cards and the first two tables is computed from Creator's LinkedInPost dataset via benchmark_within_sender() on 2026-09-09. The format guidance table is compiled from Creator's own content operations and is labelled as such.
Understand the terms
Media Post
A post carrying at least one image, document or video.
Document Post
A multi-page PDF rendered as a swipeable carousel in feed.
Dwell Time
How long a reader stays on a post before scrolling.
Truncation Point
Where LinkedIn cuts a post body and shows the see-more link.
Within-Account Median
Each account's own median taken first, so heavy posters cannot skew the benchmark.
Engagement Rate
Reactions plus comments plus shares, divided by impressions.
LinkedIn Content Playbook
How to build a visual for every post without adding a day of work.
More LinkedIn benchmarks
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