LinkedIn Text Posts vs Carousels vs Video: How to Choose the Right Format

Every few months the same argument circulates: video is winning, or carousels are dead, or text is back. Almost all of it is based on someone's own account, in their own industry, over a few weeks. The more useful framing is not which format wins but which format suits which idea — because each one is good at a genuinely different job, and the cost of producing them differs by an order of magnitude. This piece covers what each format is actually for, what it costs you in time, the official limits worth knowing, and a decision rule you can apply in about ten seconds.

What each format is genuinely good at

Forget performance rankings for a moment and think about what each format can carry that the others cannot.

Text posts

Text is the fastest path from a thought to a published post, which makes it the only format most people can sustain for years. It is best at argument, narrative and nuance — anything where the value is in the sequence of the reasoning. It is weakest at anything spatial: a process with eight steps, a comparison across four options, a structure the reader needs to see rather than read.

Text also has the lowest production risk. There is no design to get wrong, no render to wait for, and nothing to reshoot. If your constraint is time rather than ambition, text is not a compromise; it is the correct choice.

Document carousels

A document post — a PDF, slide deck or document uploaded so readers swipe through it — is the best available format for structured thinking. Frameworks, sequences, comparisons, teardowns and step-by-step processes all land better as slides than as paragraphs, because the reader can see the shape of the argument before reading any of it.

The cost is real. A good carousel needs the thinking, a structure, and design discipline — usually somewhere between one and three hours for someone who is not a designer. That is affordable occasionally and unsustainable weekly unless you have a repeatable template. The other cost is that carousels are easy to make badly: too many words per slide, a structure that is really a text post chopped into slides, or design that overwhelms the content.

Native video

Video carries things neither other format can: tone of voice, a demonstration, and the sense that there is a person behind the account. For consultants and founders whose product is effectively their judgement, that last one is commercially significant. Video is strongest for short arguments, demonstrations, and anything where seeing the thing matters more than the precise wording.

It is also the most expensive and the least forgiving. Scripting, recording, editing and captioning take real time, and a video that rambles is worse than no video at all. One practical constraint that catches people out: feed video autoplays without sound, so a video that depends on audio is invisible to most of the audience it reaches. Captions are not an accessibility nicety here; they are the difference between being watched and being skipped.

The other formats worth knowing

• Polls are cheap to produce and reliably generate votes, but votes cost the reader almost nothing and rarely produce conversation. Useful for a quick read on a question; near-useless for building a position.
• Single images work when the image is the content — a chart, a diagram, a screenshot with annotation. A decorative stock image adds nothing.
• Newsletters and articles suit long-form writing that needs to be findable later, and they behave differently from feed posts: they persist and can be referenced, which makes them useful as the home for a substantial argument you then post about.

A decision rule

One more consideration that should override the table when it applies: choose the format you will actually produce. A carousel you never finish is worth nothing, and a text post that ships beats a video that does not.

Production cost, honestly

Those are honest ranges rather than measurements, and your own numbers will differ. The point is the ratio: video typically costs several times what text does, and a carousel sits in between. Plan a cadence you can afford at those ratios rather than one that assumes you will get faster.

The limits worth knowing

LinkedIn publishes its own specifications, and those are the authority rather than any secondary write-up. Two are worth knowing because they shape how you write and produce.

The character limit for a feed post is 3,000 characters, per LinkedIn's help centre on posting updates. That page also notes something more consequential that gets far less attention: LinkedIn may limit how many posts a member or Page can make in a given period, based on behavioural and authenticity signals such as how members interact with your content and whether people block or report it. In other words, volume is not a free variable, and a sudden increase in posting frequency is not without risk.

Video specifications are published separately, in LinkedIn's help page on video specifications, including maximum duration, file size bounds, resolution range and aspect ratio. Media specifications do change and can differ between surfaces, so verify against LinkedIn's current documentation before you build a production process around an exact figure rather than trusting a secondary source — including the numbers quoted in blog posts, this one included.

Mistakes that cost the most

• Chopping a text post into slides and calling it a carousel. A carousel should have a structure the reader can see; if it reads as paragraphs with page numbers, it should have been text.
• Publishing video without captions. Feed video plays without sound. Without captions or on-screen text, most of the audience sees a silent moving image.
• Choosing a format because it is said to perform. Format follows the idea. A format chosen for its reputation produces content whose shape fights its substance.
• Over-producing the easy format and never attempting the hard one. The reverse error is equally common: committing to weekly video and quietly abandoning it in week three.
• Ignoring the first slide or first second. Both act as a second hook, doing the same job as the opening line. A weak first slide loses the reader before the substance starts.

How to test formats without fooling yourself

The temptation is to compare average performance by format and conclude that one wins. That comparison is usually invalid, because the formats are not carrying equivalent ideas — the carousels are the ones you thought worth designing, the videos the ones you thought worth recording. You are measuring your selection, not the format.

A fairer test: take one strong idea and publish it as text, then publish a genuinely different idea as a carousel, and again as video, and compare within each idea rather than across formats. Better still, publish the same idea in two formats a month apart and compare the response each produced. Our piece on calculating engagement rate sets out how to make that comparison mean something, including why impressions alone will mislead you.

Format is not a growth lever. It is a fit question: choose the one that carries this idea most clearly, at a production cost you can sustain for a year.

Whichever formats you settle on, the discipline is the same: decide the format when you plan the slot, not when you sit down to write. Zybble's LinkedIn scheduler lets you queue text, image and document posts ahead of time across the accounts you manage, and Zybble's analytics shows per-post performance so you can compare like for like. The pricing page sets out what each plan includes.

Frequently asked questions

Match the format to the job, budget honestly for what it costs to produce, work within LinkedIn's published limits, and test by idea rather than by format. Do that and the format question stops being a source of anxiety and becomes a ten-second decision. If you want the publishing handled, see how Zybble works and compare the plans.

Carousel design that holds attention

Most carousels underperform for design reasons rather than content reasons. The ideas are fine; the slides are unreadable. Four principles fix the majority of them.

• One idea per slide. If a slide needs a paragraph, it is two slides. Readers swipe; they do not read
• Set type large enough to be read on a phone held at arm's length. Design on a phone screen rather than a large monitor, because that is where it will be read.
• Use contrast rather than decoration. A limited palette — two colours plus white — reads as considered. Five colours reads as a template nobody edited.
• Make the first slide carry content. A title slide is a wasted swipe, and the swipe is the scarce resource.
• Signal progress. Numbered slides or a visible sequence tell the reader how much is left, which materially affects whether they continue.

A structural point that is easy to miss: the last slide is the one that earns the comment. Ending on a conclusion nobody can respond to wastes the format's strongest moment. End on the question you genuinely want answered, or on the step the reader should take first.

A minimum viable video setup

Video is expensive mainly because of the equipment people assume they need. It is not. The setup below produces acceptable results and costs very little.

• A phone on a stand, near a window, facing the light. Natural light from the front beats any affordable lighting kit.
• An inexpensive lapel microphone that plugs into the phone. Audio quality matters more than video quality, because viewers tolerate a soft image and abandon bad sound almost immediately.
• A fixed background with nothing moving in it. Visual distraction is more costly than visual dullness.
• A short script structure rather than a full script: the point in one sentence, three supporting beats, one closing line. Full scripts produce wooden delivery; no script produces rambling.
• Captions, always. Feed video plays without sound, and a captioned video is also accessible to viewers who are deaf or hard of hearing, which is a real part of any professional audience.

Keep the first takes rather than polishing. Most people spend four times as long editing as recording on their early videos, and the improvement in the result is far smaller than the time spent.

Accessibility in every format

This is a quality issue as much as an inclusion one, and it applies to each format differently.

• Text: avoid long unbroken blocks, and avoid conveying meaning through emoji alone. Screen readers announce emoji literally, and a line of symbols is noise to a non-sighted reader.
• Images and carousels: write meaningful alt text where the platform allows it, and never rely on colour alone to convey a distinction — roughly one in twelve men has some form of colour vision deficiency, and a chart that depends on red versus green is unreadable to them.
• Video: captions, plus on-screen text for anything essential. Auto-captioning is a starting point and usually needs correcting, particularly with product and company names.
• All formats: capitalise normally rather than using alternating case for emphasis, which some screen readers handle unpredictably.

Format and distribution: what is actually known

It is worth being explicit about the limits here, because the format debate is dominated by confident claims that the evidence does not support.

What is reasonably well established: the platform shows a post to a limited initial audience and expands distribution based on how that audience responds; early engagement matters; and each format presents a different surface for that decision, so a carousel's swipe and a video's watch time are signals of a different kind from a text post's expand.

What is not established: that any format is systematically favoured, or that the weighting is stable over time. LinkedIn does not publish its ranking system, and the format comparisons circulating online are almost always someone's own account over a few weeks. Choose formats on fit and production cost, and let your own results across a quarter — not someone else's screenshot — decide the mix.

A format mix for a solo operator

If you are one person without a content team, the question is not which formats exist but how many you can sustain without the quality collapsing. The honest answer for most people is two.

A workable default: text as the spine, plus one structured format — usually carousels — produced fortnightly from a fixed template. That gives you variety, a format suited to the ideas that need structure, and a production cost you can predict. Add video when you have a specific reason for it rather than as a permanent obligation, and treat polls as occasional research rather than a format.

The reason two formats beats three is that each additional format carries a fixed overhead: a template to maintain, a skill to keep current, and a decision to make every time you plan a slot. Three formats roughly triples the planning overhead while only modestly improving the output, and the overhead is what eventually causes people to stop planning at all.

Changing format without losing your audience

Adding or dropping a format is a change to what your audience expects, and it is worth handling deliberately rather than abruptly.

• Add formats gradually. Introduce video as one post a month for a quarter rather than as a weekly commitment, and see whether you can sustain it before you depend on it.
• Signal the change. A short post explaining that you are going to try something different converts a jarring change into something readers are watching for, and gives you permission for the early attempts to be rough.
• Never drop your strongest format to make room. If text is what your audience responds to, add video alongside it rather than instead of it.
• Watch whether the new format reaches people the old one did not. That is the only compelling reason to add a format, and it takes several posts to see.

The failure mode to avoid is format churn: adopting a new format every quarter because the last one did not transform your results. No format transforms results. Fit and consistency do, and both require staying with a decision long enough to evaluate it.

A note on reposting and recycling

Republishing an old post is occasionally worthwhile and routinely overdone. The case for it is real: most of your audience did not see the original, because distribution at the time may have been poor through no fault of the content. The case against is that repetition is visible to the small group who engage most, and those are precisely the people you cannot afford to bore.

• Republish only when the original underperformed for a distribution reason rather than a quality one — a bad time slot, a quiet week, or a period when you had a much smaller audience.
• Wait at least six months. Anything sooner is visible repetition to your regular readers.
• Rewrite the opening rather than reposting verbatim. A new hook makes it a new post for the people who saw it and removes the sense of being shown the same thing.
• Never republish something that performed well the first time. It already reached everyone it was going to.

The more reliable version of this is not republishing but relinking: referencing an older post from a new one. That gives the old piece a second life with readers who are already engaged, without asking anyone to read the same thing twice.