The B2B Video Reality Check: Are Videos Actually Working on LinkedIn?

Videos are everywhere. But are they actually working?

LinkedIn says video viewing is growing. Marketing surveys say B2B teams are investing more in it. Feeds certainly feel more video-heavy. But adoption is not the same as performance, and platform-wide growth does not tell an individual B2B company whether its next idea should become a video.

So Invisible Keyboard collected the posts and compared the formats.

The headline findings

Visible interactions means reactions + comments + reposts. Public LinkedIn pages do not provide reliable post-level impressions, watch time, leads, pipeline, or revenue, so this study does not pretend to measure them.

Does video perform better on LinkedIn for B2B companies?

In this sample, video performed better more often than it performed worse. It beat the same account's non-video median in 16 of 25 comparable company pages. The median company-level lift was approximately 19%.

That is useful evidence in video's favor. It is not evidence that video automatically wins.

Video is a multiplier. It is not a rescue plan for an idea no one cares about.

The same format produced very different results across credible B2B brands. Supabase, Datadog, OpenAI, ClickUp, and Deel recorded materially stronger median interaction levels on video during the observation window. MongoDB, Atlassian, GitLab, Dropbox, and Elastic performed better with their other formats.

A strong idea may become easier to notice, understand, demonstrate, or share when it moves. A generic idea does not become interesting simply because someone records it.

Why we compared each company with itself

Comparing an OpenAI video with a text post from a company with 2,000 followers mostly measures differences in brand strength, audience size, and existing distribution. It tells us very little about format.

Our primary comparison therefore used only company pages with at least five eligible videos and five eligible non-video posts. We calculated each company's median visible interactions for both groups, then compared the two medians. This reduces, but does not eliminate, differences between accounts.

What happened to comments and reposts?

• Video generated more median reposts in 16 accounts, tied in two, and lost in seven.
• Video generated more median comments in 14 accounts, tied in four, and lost in seven.
• Across the comparable sample, the median video received 131 visible interactions versus 108 for non-video posts.

Those results point in the same direction as the headline comparison, but they still describe visible engagement rather than commercial impact.

How confident should B2B marketers be?

Directionally confident. Statistically humble.

Sixteen wins in 25 comparable accounts is not overwhelming evidence by itself. An exact two-sided sign test returns p = 0.23. An account-level bootstrap 95% interval for the median video ratio spans approximately 0.92x to 1.38x, crossing the neutral point of 1.0x.

We also repeated the comparison with a stricter cutoff that gave every post at least seven days to mature. Video still won in 14 of 23 accounts, while the estimated median lift fell to approximately 15%. The direction survived; the size of the advantage became smaller.

Across this sample, video won more often than it lost and delivered a modest median advantage. The result varied too much by company to treat format as the main cause.

What B2B teams should do with this

1. Do not create video just because the feed contains more video. Start with an idea that benefits from demonstration, a human voice, movement, or emotion.
2. Run a within-account test. Publish at least five credible videos and compare their median performance with your own non-video baseline.
3. Measure the job of the content. Product demos may need clicks or trials. Founder videos may need qualified conversations. Customer stories may need reuse by sales.
4. Keep production proportional to the idea. A useful screen recording can beat an expensive brand film. Production quality and strategic quality are different variables.
5. Do not kill the test after one post. One strong or weak result tells you more about that post than it does about video as a channel.

Study methodology

• Sampling frame: 70 prominent, US-focused B2B software company pages were attempted.
• Eligible sample: 57 company pages published eligible posts in the observation window.
• Collection: 1,167 original organic LinkedIn posts collected on August 28, 2026.
• Primary window: posts published from May 30 through August 25, 2026, excluding the final two incomplete days before collection.
• Primary dataset: 1,009 posts, including 379 videos and 630 non-video posts.
• Within-company comparison: 583 posts from 25 pages with at least five videos and five non-video posts.
• Primary metric: visible interactions, calculated as reactions + comments + reposts.
• Summary statistic: medians, used to reduce the influence of a small number of viral posts.

We excluded reposts without original commentary, employee or partner posts surfaced inside company feeds, unavailable posts, obvious duplicates, and posts without the public data required for analysis.

What this study does not prove

• It does not establish that video caused the observed difference.
• It does not measure paid distribution.
• It does not compare LinkedIn views with views from other platforms.
• It does not prove that engagement became pipeline or revenue.
• It does not yet distinguish talking heads, demos, interviews, customer stories, or launch films. Those variables require a separate manual visual-coding pass.
• The sample overrepresents prominent B2B software brands and should not be generalized to every company or industry.

How this fits with existing B2B video research

The market is unquestionably producing and consuming more video. LinkedIn reported video viewing up 36% year over year and video creation growing twice as fast as other original formats. LinkedIn and Ipsos reported that 78% of surveyed senior B2B marketers used video. Content Marketing Institute reported that 58% of surveyed B2B marketers rated video as their most effective content type.

Those studies establish adoption, investment, and marketer belief. They do not prove that an organic video will outperform another format on a specific company's page. That is the gap this study begins to examine.

Sources

Invisible Keyboard's original dataset and calculations produced the company-level findings above. External context comes from the following published research: LinkedIn: 13 Top Tips for Compelling B2B Video Content · LinkedIn Creative Labs: analysis of 13,000 B2B video ads · LinkedIn and Ipsos: 2025 B2B Marketing Benchmark · Content Marketing Institute: 2025 B2B Content Marketing Benchmarks · Vidyard: 2025 Video in Business Benchmark Report · Wistia: 2026 State of Video Report

Platform-produced studies are labeled as platform research. Findings from paid-video studies are not presented here as proof of organic-video performance.

Frequently asked questions

Does video perform better than text or images on LinkedIn?

In this study, video beat the same company's non-video baseline in 16 of 25 comparable B2B software accounts, with a median company-level lift of 19% in visible interactions. It lost in nine accounts, so video did not reliably outperform other formats for every company.

What does visible interaction mean?

Visible interactions are the sum of public reactions, comments, and reposts on a LinkedIn post. They do not include impressions, watch time, profile visits, leads, pipeline, or revenue.

How many LinkedIn posts were analyzed?

Invisible Keyboard collected 1,167 organic posts from 57 B2B software company pages. The primary analysis used 1,009 posts, and the strict within-company video comparison used 583 posts from 25 company pages.

Should every B2B company start making videos?

No. B2B companies should test video when the idea benefits from demonstration, voice, movement, or emotion, then compare results with their own baseline. The format should serve the idea and commercial goal, not replace them.

Need the content system behind the test?

Invisible Keyboard turns the work, customer conversations, product decisions, and market POV already inside a company into founder and team content, then runs the publishing system around it.See how Invisible Keyboard works