LinkedIn Video Ads for B2B: Length, CTA & Cost-Per-Lead That Works

A RevOps tool I worked with ran two LinkedIn video ads for the same offer over a month, against identical targeting (Sales Operations + RevOps titles, 200-1000 headcount, US) and budget. Version one was a glossy 75-second sizzle reel: founder to camera, soft music, "we believe sales should be effortless." Version two was a 14-second silent clip, one line of white text on a dark frame, then a screen recording. The sizzle reel returned a cost-per-lead north of $180. The 14-second clip landed under $70 and held there. Nothing about the product changed; the structure did.

That gap is the whole story of B2B video on LinkedIn. The muted feed and a senior, skeptical buyer punish video built like a TV spot and reward video built for a silent thumb-stop. Below is how pacing, length, and CTA logic fit LinkedIn specifically, plus what those decisions do to your cost-per-lead.

Two failure archetypes you can name on sight

After enough audits you stop seeing unique mistakes and start seeing the same two films.

The 90s Brand Film. A high-production piece that treats LinkedIn like a cinema: it builds, earns its payoff at the end, and leans on a voiceover. The problem is mechanical. Average view duration on a cold LinkedIn audience rarely clears 5-6 seconds, so the customer quote at 0:71 is never seen, and the VO is muted anyway.

The "We Help Companies..." Open. The first frame is a logo card or a sentence describing the advertiser instead of the viewer. "We help B2B teams scale faster." A director scrolling between meetings has no reason to watch that, and it spends the only window you get talking about yourself. The format should follow the funnel stage you target, but neither archetype survives a cold feed regardless of format. Both fail for the same three reasons, none of them your product.

Why B2B LinkedIn video ads break in a muted feed

Autoplay runs silent. Video starts as the ad scrolls into view, no sound, unless the viewer taps to unmute, and on LinkedIn most never do. If your first frames need a voiceover, you have lost the room. The implication runs deeper than "add captions": a muted feed also changes which campaign objective you should pick, which I will come back to.

The feed is a thumb-stop economy. You compete with a colleague's promotion, a layoff post, and a hot take on hybrid work. The opening frame gets one scroll-pause to earn the next two seconds. The window shifts depending on where the ad runs, and LinkedIn's is among the shortest.

The buyer is senior and has seen every promise. LinkedIn's targeting is its real product: you can buy directors and VPs by title, seniority, function, and company size. That precision is also the trap, because those people have read "transform your workflow" a thousand times. Vague aspiration reads as noise; a specific, falsifiable claim reads as someone who knows their job.

Length and the time-to-hook number

On LinkedIn, completion rate falls off a cliff after the first several seconds, so treat every second past 15 on a cold audience as a tax paid in dropped viewers. Working ranges I trust:

• Cold prospecting: 6-15 seconds. One idea, one hook, one visual payoff. You are buying a name and a retargetable view, not a decision.
• Consideration and retargeting: 15-30 seconds. People who know you will sit through a demo or a real proof point.
• Long-form on cold traffic: almost never. If you have a strong 90-second video, cut a 10-second teaser as the ad and host the full version on the landing page.

The metric that beats total length is time-to-hook: the viewer needs to grasp what this is and why it touches their job before the third second. Front-load the payoff, because saving it for a reveal on LinkedIn means saving it for nobody.

Real B2B hooks, written out

"Lead with the problem" is useless without examples. Here are three opening lines I have shipped, by category, each built to read as legible text on a muted, half-glanced frame.

RevOps tool. Plain text: "Your reps rebuild the same deck 40 times a quarter." Then cut to a screen recording of the deck assembling itself. No logo, no founder, no music. The number does the work; a Sales Ops leader recognizes that waste instantly.

Recruiting service. Text over a calendar full of declined slots: "23 days to fill a senior backend role. Most of it is sourcing you could skip." Then a shortlist populating. It names a metric the hiring manager is judged on, not a benefit you invented.

Fintech / spend management. A frozen expense dashboard: "Finance closes the books 6 days late because approvals live in Slack." Then approvals flowing inside one tool. The pain is operational and dated to a specific person's month-end.

Notice what none of them do: name the company, show a logo card, or use a verb like "transform." For more angles, a set of hook templates gives you enough patterns to fill a test, and the broader script framework carries the same beats into longer cuts. Keep body and CTA fixed, vary only the opening line, and you have a clean test of the highest-leverage variable.

Captions carry the message, so write them to be read

Sound is off by default, so the on-screen text is the ad. Burn it in rather than trusting auto-captions, so the hook is legible the instant the frame loads. The craft point people miss: a spoken line and a readable caption are different lengths. "What if there were a faster way to handle your team's reporting" becomes "Reporting that builds itself." Cut filler, keep nouns and verbs, and leave the bottom ~15% of the frame clear so LinkedIn's headline and CTA button do not land on your words. The mechanics of styling and timing burned-in text carry across placements, but on LinkedIn they are load-bearing.

The CTA decision that depends on your objective

This is the part most B2B guides skip, and the one thing genuinely specific to LinkedIn. Your CTA logic is tangled up with two LinkedIn choices: campaign objective, and whether you capture leads on-platform or send the click off.

A muted 14-second clip watched by a VP who learned your name eight seconds ago has not earned a "Book a demo." Match the ask to the audience, and the objective to the ask:

• Cold, brand awareness or video views objective: ask for nothing but the next view or click. A teardown, a benchmark, a short tool. You are buying a retargetable audience, not a meeting.
• Cold-to-warm, lead generation objective with a LinkedIn Lead Gen Form: the form pre-fills from the member's profile, so a senior buyer hands over a work email in two taps without leaving the feed. It is the highest-converting capture for cold-to-warm video on LinkedIn because it removes the off-platform click the muted, distracted viewer will not make.
• Warm and retargeted, website conversion objective: now an off-platform click to a demo, trial, or pricing page is justified. They have context, so the friction pays for itself.

The trade-off is real. Lead Gen Forms get you cheap, high-volume leads of lower average intent, since the two-tap submit is frictionless for tire-kickers too. Off-platform conversion gets fewer, higher-intent leads at a higher cost-per-lead. For most cold B2B video, start with the form, then retarget the engaged segment off-platform with the demo ask. Whichever you pick, use one CTA, match the on-screen text to the LinkedIn button and to the form or page headline, and keep a stage-appropriate CTA rather than defaulting to the close.

What LinkedIn video actually costs

Here are directional ranges that vary by industry, geography, and competition. Treat them as the order of magnitude to budget against, not a quote.

• CPM: LinkedIn is expensive media. Expect roughly $30-$80 CPM for tightly targeted senior B2B audiences, higher in crowded categories like enterprise software or finance. It is several times the cost of Meta because you are paying for the title-level targeting.
• CPC: commonly in the $6-$12 band for B2B, climbing with seniority and competition.
• CPL via Lead Gen Forms: often $30-$120 for mid-funnel offers (guides, benchmarks, webinars). The form's low friction keeps this down.
• CPL for off-platform demo requests: realistically $150-$400+ on cold-to-warm traffic for considered B2B purchases. Higher intent, higher price.

Two things move these numbers more than your bid. First, the hook: a clip that holds viewers past three seconds cuts effective CPM, because LinkedIn rewards engagement with cheaper delivery. Second, the objective and capture choice above. A structurally tight video pointed at a Lead Gen Form is how the RevOps example got under $70 while the sizzle reel sat at $180 on identical media cost.

Why iteration is the lever

When a B2B video underperforms, the reflex is to re-edit: new music, tighter color, a fresh voiceover. The data almost never supports it, because the drop-off lives in the first three seconds. The lever is the hook, not the craft. So structure testing around it: hold body and CTA constant, produce four or five openers that differ only in the first line (number, pain, contrarian, question), and run them against the same audience until each reads on view-through and click. The discipline of changing one variable at a time keeps results legible.

The catch is throughput. This is cheap when each variant costs minutes and slow when every cut is a manual edit, which is why most teams ship one hook. Closing that gap is what faster iteration buys you.

FAQ

Should I use LinkedIn Lead Gen Forms with video?

For cold and cold-to-warm video, usually yes. The form pre-fills from the member's profile, so a senior buyer submits in two taps without leaving the feed, which matters because a muted-feed viewer rarely follows an off-platform link. Expect higher lead volume at lower average intent, and reserve off-platform conversion clicks for warm, retargeted audiences with context.

What does a LinkedIn video ad cost for B2B?

Directionally: CPMs run roughly $30-$80 for tightly targeted senior audiences, CPCs around $6-$12. Lead Gen Form CPLs often land $30-$120 for mid-funnel offers, while off-platform demo-request CPLs are realistically $150-$400+ on cold-to-warm traffic. These shift with industry and geography, and a stronger hook lowers effective CPM by improving engagement.

Single image, video, or document ads for B2B?

Single-image ads are cheapest to test and fine for direct-response to warm audiences. Video wins for cold prospecting when you need to show a product in motion and earn a thumb-stop. Document ads (a swipeable PDF) suit educational, high-information offers like frameworks or reports. A common sequence: video to introduce, document to educate the engaged, single image to retarget with the offer.

If the limit on all this is producing enough hook variants to run a real test, that is the narrow thing Aitachyon handles: generate several captioned cuts of a LinkedIn video ad quickly, so testing five openers costs about what testing one used to. It will not pick your objective, but it makes the iteration above cheap enough to actually do. Start here.