Most landing-page videos hurt conversion. Some lift it dramatically. The difference is roughly six conditions — all of which feel obvious in retrospect, almost none of which most teams test for, and exactly one of which most teams get backwards.
The short answer: video lifts conversion when it answers the buyer’s most-asked question faster than text can. It kills conversion when it asks for time the buyer hasn’t decided to give yet.
The six conditions under which landing-page video lifts conversion
1. The product is genuinely visually demonstrable.
Video helps for products you can show — a UI flow, a physical product in use, a transformation before-and-after, a process. Video hurts for products that are conceptually simple but visually abstract (financial services, B2B SaaS for invisible systems, professional services). For a conceptually simple product, three bullet points often outperform a 90-second explainer. The buyer is asking “is this for me?” — they don’t need 90 seconds to answer that.
2. The buyer’s #1 question is “how does this actually work” — not “what is it” or “why this one.”
Video shines when the buyer has already decided what the product is and is trying to understand the mechanism. It under-performs when the buyer is still deciding whether the category is right for them. Different funnel stages need different content; treating video as a one-stage-fits-all asset is the most common production mistake.
3. The video is short — ≤60 seconds at the fold, ≤120 seconds anywhere on the page.
Hero-fold videos over 60 seconds have measurably lower completion than they’re given credit for. Above 90 seconds, completion typically falls below 30%. The most-cited “engagement metric” in agency pitch decks — “average watch time” — is misleading; what actually correlates with conversion is the *percentage of viewers who complete the video and then act on the page*. Short, captioned, conclusive video does this; long cinematic-feel video does not.
4. The video is silent-by-default and works without sound.
Roughly 85% of landing-page video plays are muted. Audio-led video that depends on a voiceover to make sense is a video that does not make sense to 85% of its audience. Captions, on-screen text, and motion clarity have to carry the message. The exception: some product categories (music, audio, food) where sound is part of the product itself.
5. The page lets the user skip the video entirely.
The buyer should be able to scan past the video and convert without watching it. If the call-to-action button is gated behind a play sequence, or the answer to “is this for me” is locked inside the video and nowhere in the page copy, conversion drops. Treat video as a confirmation layer for the scanner, not a replacement for clear copy.
6. The video loads under 2 seconds and does not blow LCP.
Largest Contentful Paint is the single biggest performance metric that affects landing-page conversion. A 4 MB hero video that takes 3+ seconds to start makes the LCP tank and drags conversion with it. Poster frame instant, video lazy-loaded, file size under 2 MB for the hero, MP4 with H.264 (or H.265 with H.264 fallback for older browsers). On mobile, autoplay with sound is explicitly disallowed by Safari — don’t fight the platform.
The five conditions under which video hurts conversion
Autoplay with sound. Always. There is no scenario where unmuted autoplay improves landing-page conversion. It interrupts the buyer’s environment (often a quiet office, sometimes a public space), it raises the cost of staying on the page, and it signals “this brand does not respect the user’s context.” This is the single mistake we see most often, and the one most resistant to client feedback.
Cinema-style brand film with no product information. The 90-second mood reel with slow-motion shots, an inspirational voiceover, and the product visible for four seconds total. Beautiful in a portfolio; converts badly on a landing page. Beauty without information signals “we have nothing concrete to show.” For brand-awareness campaigns at the top of the funnel, mood reels can work; on a landing page asking for a click, they almost never do.
Hero video is the only path forward. No bullets, no value propositions, no benefit copy on the page — just the video. Scanners bounce. Skimmers bounce. The video has to *complement* the copy, not replace it.
Video weight is over 4 MB without proper lazy-loading. If the page can’t render fast because the video is in the critical render path, the page loses the buyer before the video even starts.
The page uses video where text would convert better. Some pages just don’t need video. Pricing pages, comparison pages, technical specification pages — the buyer is scanning for information, not waiting for a story.
The single biggest production mistake
Producing one hero video and treating it as a universal asset. Different funnel stages need different video. The explainer that converts top-of-funnel traffic is the wrong asset for the bottom-of-funnel landing page; the product demo that converts in nurture emails is the wrong asset for a cold-traffic hero. We recommend planning video in tiers: one hero loop or short explainer for the landing page, derivative product demos for mid-funnel pages, social cuts for traffic acquisition. One commission, three to five outputs.
The three-question test before commissioning landing-page video
- Can the same point be made in three bullets and an image? If yes, video is decoration, not utility. The product probably doesn’t need it.
- Will the video work on mute with captions only? If no, the video will fail 85% of its viewers.
- Will the video load in under 2 seconds and run under 60? If no, the video will tank the page’s Core Web Vitals or be skipped before it lands its point.
Three yeses, commission the video. Two yeses, redesign the brief. One yes or zero, find a non-video solution.
When video is a feature vs decoration
The reliable threshold: video is a feature when the buyer’s most-asked question is “how does this work” and the product is visually demonstrable. Otherwise, video is decoration — sometimes pleasant, sometimes neutral, often harmful, almost never the highest-impact thing the page could be doing with that real estate.
If you have already produced a hero video and are not sure whether it is helping or hurting, the answer is testable: run a two-week A/B test with and without the video on the hero. The lift or drop will be visible inside seven days for any page with reasonable traffic. We have done this for clients and the result has surprised us in both directions — but we have never seen video survive the test on a page where the product was conceptually simple or visually abstract.
Common questions
Does video on a landing page always increase conversion?
No — video lifts conversion under specific conditions and hurts it under others. The reliable lift happens when the product is visually demonstrable, the buyer’s main question is ‘how does this work,’ the video is under 60 seconds, runs silent by default with captions, can be skipped, and loads under 2 seconds. When those conditions are missing — particularly autoplay with sound, cinema-style brand films, or videos that block the call-to-action — landing-page video typically reduces conversion. The most reliable way to know which side your page is on is a two-week A/B test.
How long should a landing page hero video be?
Under 60 seconds at the fold; under 120 seconds anywhere else on the page. Completion drops sharply above 60 seconds, and the percentage of viewers who watch to the end and then act on the page — the metric that actually correlates with conversion — typically falls below 30% for hero videos over 90 seconds. Short and conclusive beats long and cinematic on a landing page. Save the longer cuts for product pages, sales decks, and nurture emails where the buyer has already committed time to learning.
Should landing page video autoplay with sound or muted?
Muted, always. Roughly 85% of landing-page video plays are muted, and unmuted autoplay measurably drops bounce-rate metrics and conversion. Major browsers (Safari and Chrome on mobile especially) block autoplay-with-sound at the platform level anyway. The right pattern is silent autoplay by default, captions on by default, an obvious mute/unmute control, and the audio version available only after a deliberate user click. Audio-led video that requires sound to make sense is failing 85% of its audience by design.
Will video on the homepage hurt my Core Web Vitals score?
It can — if implemented carelessly. The risks: a heavy video file in the critical render path tanks Largest Contentful Paint; an autoplay video on first load delays Time To Interactive; a video that takes its own layout shift after load hurts Cumulative Layout Shift. The fix is mechanical: poster frame served instantly (under 50KB), video file lazy-loaded with proper preload hints, video file size under 2 MB for hero use, and the player loaded after first paint. Done right, video has no measurable impact on Core Web Vitals; done wrong, it can drop the page’s performance score by 10–20 points.
Is animated SVG or Lottie a better fit than video for landing pages?
For most landing-page hero work, yes. Animated SVG and Lottie animations load faster, render sharper at any size, work as vector graphics on retina displays, are scriptable from JavaScript, and weigh kilobytes rather than megabytes. Video still wins when the content is photographic, when audio is part of the message, or when the motion is complex enough that hand-animating it as a Lottie is more expensive than filming it. We typically recommend a tiered approach: Lottie for the hero motion, video for the longer-form explainer further down the page, social cuts derived from the video for outbound channels.
Need a hero video that actually lifts conversion?
We design landing-page video, motion, and Lottie animations built around the six conditions that move the metric.