Most guides on how to upload video to a LinkedIn post stop at "click the video icon." That's the wrong place to stop.
The upload is mechanical. What LinkedIn does with your file in the hours after publish is where distribution is won or lost.
What does LinkedIn do with your video file in the first 72 hours?
The moment you hit publish, LinkedIn queues your file for re-encoding. It converts your original upload into multiple resolution variants to serve different connection speeds and devices. This process usually completes within minutes, but the distribution window opens immediately.
In the first few hours, LinkedIn shows your video to a small, controlled slice of your network. It measures two things above everything else: how long people watch, and how many watch to the end. These signals, not likes or comments, determine whether the algorithm extends reach to second-degree connections and beyond.
The implication is direct: if your video loses the audience in the first 10 seconds, the algorithm reads that as a quality signal and stops pushing it. The first distribution window is effectively your only one. A post that underperforms in the first few hours rarely recovers, regardless of what you do afterward. For a deeper look at how these signals interact, Posting Video to LinkedIn: What Actually Drives Reach covers the mechanics in detail.
Which format constraints actually matter, and which ones don't?
LinkedIn's official specs list a range of accepted formats, frame rates, and bitrates. Most of it is noise for practitioners.
The constraints that genuinely affect your workflow are three. First, file format: MP4 encoded with H.264 is the safest choice across every device and upload path. MOV works, but re-encoding artifacts are more common. Second, file size: the 5 GB ceiling is generous, but large files slow the re-encoding queue and can delay your post going live. Keep exports under 500 MB for anything under 5 minutes. Third, aspect ratio: square (1:1) and vertical (4:5 or 9:16) formats take up more screen real estate in the mobile feed than 16:9 landscape. More screen space means more passive watch time before a scroll.
Resolution caps at 4K, but 1080p is where the quality-to-file-size trade-off is optimal for most LinkedIn content. Anything above that is re-encoded down anyway.
The constraints that don't matter in practice: exact frame rate (24, 25, 30 fps all behave identically in the feed), audio codec (AAC is standard and universally accepted), and minimum bitrate thresholds (LinkedIn's encoder compensates for low-bitrate sources).
Why does native upload outperform a YouTube link in LinkedIn distribution?
The distribution gap between native video and a YouTube link is not a rumour. It is a structural feature of how LinkedIn ranks feed content.
LinkedIn's feed ranking penalises posts that route users off-platform. A post containing a YouTube URL is treated as an external link post, not a video post. It competes in a different, lower-priority content category. The video thumbnail may render, but the distribution logic is the same as any other link post: LinkedIn has no incentive to push content that sends its users to Google's platform.
Native video keeps the viewer on LinkedIn. Every second of watch time is a signal LinkedIn can measure and attribute. With a YouTube link, LinkedIn gets nothing after the click. That asymmetry explains the reach gap.
The practical rule: if the video lives on YouTube for SEO reasons, fine. But post the native file to LinkedIn separately. Don't cross-post the link and expect equivalent reach. The two channels serve different distribution goals and should be treated as independent posts.
How should you read video post performance differently from text posts?
Text posts and video posts generate fundamentally different signal sets. Treating them the same way leads to bad editorial decisions.
For a text post, impressions and engagement rate (reactions + comments divided by impressions) are reasonable proxies for performance. For a video post, those metrics are secondary. The primary signals are completion rate (what share of viewers watched to the end) and average watch time (how many seconds the median viewer spent).
A video with 10,000 impressions and a 15% completion rate is performing well. A video with 40,000 impressions and a 4% completion rate is being shown to people who immediately scroll past — the algorithm inflated impressions but the content didn't hold attention. The second scenario often looks better in a vanity dashboard and is actually worse for your account's long-term distribution health.
This is where DSB Intelligence's Insight Narrator is useful: it reads video post data through the lens of watch-time signals rather than aggregating all post types into a single engagement rate, which would obscure the distinction entirely.
For context on how posting cadence interacts with video performance, How Often to Post on LinkedIn: Daily Is Backfiring is worth reading alongside this.
Can you post a 10-minute video on LinkedIn? The honest answer on length
Yes, LinkedIn allows videos up to 10 minutes on standard posts. The technical ceiling is real.
The practical ceiling is lower. Completion rate drops sharply as video length increases, and completion rate is the signal LinkedIn uses to extend distribution. A 90-second video with a 40% completion rate will outperform a 9-minute video with a 6% completion rate in almost every distribution scenario.
The honest guidance: match length to the content's actual density. A product demo that needs 7 minutes should be 7 minutes. A talking-head opinion piece that could be 90 seconds should not be padded to 5. LinkedIn's audience is professional and time-constrained. They will not watch out of politeness.
If you have genuinely long-form content (webinar recordings, conference talks), consider cutting a 2-3 minute highlight reel as the native LinkedIn post, and linking to the full version in the first comment rather than the post body. This keeps the post body free of external links while still giving interested viewers a path to the full content. See also LinkedIn Articles vs Posts: What Actually Builds Authority for how long-form content fits into a broader content strategy.
What are the three things to do before you hit publish?
Getting the upload right is a pre-publish checklist, not an afterthought.
-
Add captions. LinkedIn allows SRT file upload at publish time. Use it. A significant portion of the LinkedIn feed is browsed on mute, particularly on mobile during commutes or in open offices. Captions are not an accessibility add-on — they are a reach multiplier. If you don't have an SRT file, burn captions directly into the video export.
-
Write the first line of your caption as a standalone hook. LinkedIn truncates post captions in the feed after two or three lines. The text visible before "see more" is the only copy most viewers will read before deciding whether to watch. Front-load the reason to care, not the context.
-
Check your posting window relative to your last video post. Spacing video posts gives the algorithm time to exhaust each post's distribution potential before the next one competes for the same audience segment. Posting two videos in quick succession splits the attention pool and typically reduces the reach of both. The same logic applies to your overall posting cadence, which How Often to Post on LinkedIn: Daily Is Backfiring addresses directly.
Now what?
Four actions you can take today:
- Re-export your next LinkedIn video as a square or vertical 1080p MP4 under 500 MB, and upload it natively rather than linking from YouTube.
- Generate or commission an SRT captions file before your next publish — most transcription tools produce one in under 5 minutes.
- Pull completion rate and average watch time on your last three video posts and compare them to your text post engagement rate. If you've been optimising for the wrong metric, you'll see it immediately.
- Track your next video post's performance in the first 6 hours specifically — that window tells you more than the 30-day total.
If you want a cleaner read on how your video posts are actually performing relative to your text and document posts, try DSB Intelligence free and let the Insight Narrator surface the signal you've been missing.