Most B2B video marketing goes wrong before anyone writes a word. The common sequence is that a team agrees it needs video, commissions a film of two or three minutes about the company or the product, and only then asks where it will run. By that point the length is fixed, the aspect ratio is fixed, and the decision about whether anyone hears the audio has been made by someone who was never in the room: the platform.
The order is backwards, and it is expensive in a specific way. The film is usually good. It is simply longer than any paid placement is built to carry, so it ends up on a website page few people reach and in a sales deck where it is rarely played. The budget bought an asset and no distribution.
The video most B2B teams commission has nowhere to run
Look at what the platforms actually say about length, in their own current documentation rather than in a benchmark report.
On YouTube, a skippable in-stream ad can be skipped after 5 seconds, and Google charges the advertiser after 30 seconds of viewing or the full completion of the video, whichever comes first. A bumper ad is 6 seconds or shorter. A non-skippable in-stream ad is 60 seconds or shorter. On LinkedIn, video ads are permitted to run from three seconds to 30 minutes, and LinkedIn recommends 15 to 30 seconds so an ad can qualify for all possible placements across the Feed and the Audience Network.
Data table
| Measure | Value |
|---|---|
| seconds before a skippable YouTube ad can be skipped | 5 |
| seconds, the maximum bumper ad | 6 |
| seconds, top of the range LinkedIn recommends | 30 |
Set those against a film of two or three minutes. There is no paid placement in either system that is designed to deliver it. The 30 minute ceiling on LinkedIn looks like permission. It is better read as an upper bound on what the system accepts, not a statement about what it distributes well. The recommendation sitting inside that ceiling is the number that reflects how the placements behave.
This is not an argument that long video is worthless in B2B. A recorded customer session or a technical walkthrough can run twenty minutes and earn every one of them, because someone chose to watch it after already deciding they had a problem. The mistake is assuming one asset can do both jobs, and then buying the expensive one first.
This keeps happening to competent teams, and the reason is structural rather than careless. The video is commissioned as a project, with a brief, a production partner and a delivery date. The distribution is planned later, as a campaign, often by different people working to a different calendar. Nothing in that sequence forces the two conversations to meet, and by the time they do, the expensive decision has already been taken. The film exists, so the campaign is built around what exists rather than around what would work.
Work backward from the placement to the length
The practical version of this is a sequencing change, not a craft change. Decide where the video runs before deciding what it is. The placement fixes three things at once, and all three are hard to retrofit.
Data table
| Placement | Seconds |
|---|---|
| YouTube skippable, before the skip option | 5 |
| Bumper ad, maximum | 6 |
| LinkedIn recommended, low end | 15 |
| LinkedIn recommended, high end | 30 |
| YouTube non-skippable, maximum | 60 |
It fixes the length, as above. It fixes the frame, because a feed placement is competing with a thumb, and a landscape crop uses less of a phone screen than a vertical one. And it fixes whether sound exists at all, which is the constraint teams underestimate most, because LinkedIn video autoplays with the sound muted, and captions are a separate SRT file.
That last point changes what a script even is. If the opening seconds carry the message only in the voiceover, the message does not arrive. The first frame has to state the subject in text, which is closer to writing a headline than to writing a narration. Teams who treat the caption file as an accessibility afterthought are shipping a silent film with the intertitles bolted on at the end.
A useful discipline: before commissioning anything, write down the placement, the number of seconds it gives you, and the sentence that has to land inside them. If that sentence cannot be written, the video is not ready to be made, and no production budget will fix it.
Formats by stage, because the committee does not watch one video
The second structural problem is different from the first, and it is often mistaken for it. A B2B purchase is made by a group. The person who first notices the problem, the people who assemble a shortlist, and the person who signs the contract are in different moments and want different things. One film cannot serve all three, however well made it is.
Splitting by moment is what makes a modest budget go further, and it is the honest answer to the production reality question. Our view is that three plain assets aimed at three moments tend to do more than one polished asset aimed at everyone, because polish is rarely the binding constraint. Relevance usually is.
Early, when the problem is still being named, the job is recognition. Short, specific, and about the reader’s situation rather than about the product. This is the material that suits the 6 to 30 second placements, and it is where paid distribution earns its keep, because the buyer is not yet looking for you by name.
In the middle, when a shortlist is forming, the job is evidence. A product walkthrough that shows the actual interface, a real workflow completed end to end, a customer explaining the problem in their own words. Length matters less here because attention is voluntary, and this is the footage that a champion forwards internally. It belongs on the site and in the sales thread more than in a feed.
Late, when the decision is being defended to people who were not in the demos, the job is reassurance. Short clips that answer the objection the finance or security reviewer will raise. They cost little to produce, they are the ones we most often find missing, and they are the ones a champion replays inside the account.
In practice that set is unglamorous. A 20 second piece naming the problem, shot well enough to look deliberate. A five minute screen recording of the product doing the thing it is bought for, with a real dataset rather than a demo one. Two or three 30 second answers to the questions that come up in every procurement conversation, delivered by whoever actually answers them on calls. None of that needs a crew. All of it needs someone who knows what the buyer asks, which is why the constraint is usually access to the sales conversation and not access to a camera.
Distribution beyond YouTube follows from the same logic. The two surfaces ask for different things. On YouTube the viewer chose to be there and can skip after 5 seconds, so the job is to be worth continuing. On LinkedIn the video arrives muted in a feed, so the job is to be legible without sound. The same asset does not belong in both, and treating “we put it on YouTube” as distribution is how a video ends up with a view count and no audience. The rest of the demand generation programme should be pulling these assets into sequences and follow-ups rather than leaving them on a channel page.
Judge B2B video on what it moves
Views are the wrong scoreboard, and not because they are vanity. They are a measurement problem. On placements where video autoplays, a view is substantially a property of the placement’s behaviour rather than of the video, so a rising view count can mean the media buy widened and nothing else. Comparing two videos on views compares two media buys.
Better questions, in rough order of how much they cost to answer. Did the people who watched come back? Did watching change what they did next, or did the same people convert who would have converted anyway? Does the sales team report that a specific asset is being forwarded or asked about? Is a named account in the pipeline after the campaign that was not in it before?
None of those are as clean as a view count, which is precisely why the view count survives. The signal we trust most is also the least automated: sales telling you which clip an account keeps bringing up. That is worth building a feedback habit around, because it points at the asset to make more of.
The second trap is comparing a number across platforms as though it meant the same thing. Each platform defines a view in its own terms, so a report placing YouTube views beside feed views beside website plays is adding up events that were never the same thing. If a summary has to combine them, say what each one counts underneath the total. Otherwise the total gets read as a single quantity, and eventually someone plans against it.
Set the completion expectation honestly too. On a skippable in-stream placement, billing begins after 30 seconds of viewing or completion, so a video shorter than 30 seconds is charged only on a full watch. That is a budgeting fact worth knowing before the length is decided, and it is another reason the placement belongs at the start of the conversation rather than the end.
FAQ: B2B video marketing
How long should a B2B video be?
It depends entirely on where it runs. For paid feed and in-stream placements, LinkedIn recommends 15 to 30 seconds so an ad can qualify for all possible placements, and YouTube’s bumper format caps at 6 seconds. For material someone chooses to watch, such as a walkthrough or a recorded session, length is not the constraint and relevance is.
What is the maximum length of a LinkedIn video ad?
LinkedIn permits video ads from three seconds to 30 minutes. The maximum is not a recommendation. LinkedIn separately advises 15 to 30 seconds so an ad can qualify for all possible placements across the Feed and the Audience Network.
When can a viewer skip a YouTube ad?
On a skippable in-stream ad, the viewer gets the option to skip after 5 seconds. That is the window in which the subject has to be established, whatever the total length of the video.
Do I need captions on B2B video?
On any placement that autoplays muted, treat them as essential. LinkedIn documents that video ads autoplay with the sound off and takes captions as a separate SRT file, so a video whose message lives only in the voiceover arrives blank. That is our recommendation rather than a platform rule, and it is the cheapest thing on this list to get right.
Should B2B video go on YouTube or LinkedIn?
Both, but not the same asset. YouTube is a discovery and search surface where attention is chosen, which suits longer explanatory material. LinkedIn is an interruption surface where the message has to survive being muted, which suits short and specific pieces.
How much should we spend on B2B video production?
Less on the first asset, and across more of them. Three modest videos aimed at three buying moments usually do more than one expensive video aimed at everyone, because production value is rarely why the first one underperformed.
What should the first video be?
Usually the middle-of-funnel one: a genuine product walkthrough showing a real workflow. It is the asset sales tend to reach for, it costs little, and it shows quickly whether the story holds up when someone watches the product work.
How do we measure whether B2B video is working?
Not on views, which largely track how wide the media buy was on autoplay placements. Look at whether watchers return, whether a named account entered the pipeline after the campaign, and whether the sales team reports specific clips being forwarded or asked about inside accounts.
Is a company brand film worth making?
Sometimes, but rarely first, and rarely with the distribution budget attached to it. It earns its place on a website, a recruitment page or an event backdrop. It is a poor fit for paid placements that are built around a handful of seconds, and buying it first is what leaves a programme with an asset and no audience.
COSEOM Team
COSEOM is a multilingual B2B growth agency. Since 2008 the COSEOM team has run international SEO, generative engine optimization, paid media and demand generation programs for B2B software and technology companies, in English, German, Spanish and Portuguese. This account publishes our research, playbooks and product guides. More about COSEOM