VSL player features: what a direct-response video player must do

A VSL player is a video player built to sell one video to visitors who didn't ask to see it. Beyond play and pause, it has to start instantly on weak mobile connections, turn a muted start into a listening viewer, keep the story in order, reveal the offer at the pitch, and connect every view to revenue. Those five jobs decide which features matter.

What makes a VSL player different from a normal video player?

Most video players are built for libraries. A viewer picks a video, skips what bores them, and gets offered the next one. That's the right design for a course platform, a help center or YouTube.

A VSL is the opposite case. The visitor came from an ad, the video is long, it was written as one continuous argument, and the business outcome is a click on an offer that appears near the end. That flips almost every default:

Library player vs VSL player

Six decisions, and the default each player makes

Library player

• Start: click to play
• Sound: on after the click
• Seeking: full scrubbing
• The end: often suggests the next video
• Success: views and watch time
• Data: plays and minutes

VSL player

• Start: instantly, muted, with an unmute prompt
• Sound: off until the visitor taps, then restart at 0:00
• Seeking: optional, often hidden on long VSLs
• The end: the offer is already on the page
• Success: pitch reach, CTA clicks, revenue per visitor
• Data: events tied to each click and sale

The five jobs of a VSL player

Every feature in this guide does one of five jobs. When you compare players, ask which job a feature serves and which number proves it's working. Open each job in turn:

From the first frame to proof of the sale

Each job: the features that do it, the number that proves it works, what goes wrong without it

1. 1Start
2. 2Sound
3. 3Story
4. 4Offer
5. 5Proof

1Start

What happens: Adaptive streaming, a CDN and a light poster start the video instantly, even on weak mobile connections.

Why it matters: A VSL can lose a visitor before it shows a single frame. Number to watch: time to first frame.

What goes wrong:

• Spinner, black screen, bounce.

Read more: VSL Page Speed: Make Video Pages Load Fast on Mobile

2Sound

What happens: Muted autoplay, an unmute prompt, a restart at 0:00 and captions turn a muted start into a listening viewer.

Why it matters: Browsers block autoplay with sound for first-time visitors. Number to watch: unmute rate.

What goes wrong:

• Visitors watch muted, then leave.

Read more: VSL Autoplay: Muted Autoplay That Earns the Unmute

3Story

What happens: Seek control, the progress bar, the pause screen, resume and speed keep the story in order.

Why it matters: A VSL is written as one continuous argument, and every visible control hands the viewer a decision. Number to watch: retention at 1 minute and at the pitch.

What goes wrong:

• Viewers skip ahead, stall or quit.

Read more: VSL Progress Bar: Standard, Hidden or Smart?

4Offer

What happens: A timed CTA reveal shows the button and the offer block at the pitch.

Why it matters: A buy button that's visible from the first second invites people to skip to a price they haven't been sold on. Numbers to watch: pitch reach, CTA click rate.

What goes wrong:

• Price shock, or a button nobody sees.

Read more: Delayed CTA Button: When to Show the Buy Button

5Proof

What happens: Pixels, server events, click IDs and reports connect every view to revenue.

Why it matters: A player that can't tell you which ad produced a sale is a player you can't scale. Numbers to watch: revenue per visitor, EPC.

What goes wrong:

• You can't tell which ad paid.

Read more: VSL Tracking: Pixels, CAPI and Sale Attribution

Protection sits across all five. It doesn't sell anything, but a copied VSL can cost you the offer.

Fast start: adaptive streaming and a CDN

A VSL can lose a visitor before it shows a single frame. The fix lives in how the video is delivered. A plain MP4 is one file at one quality, so a phone on one bar of signal has to pull the same heavy file as a laptop on fiber.

HTTP Live Streaming (HLS), documented in RFC 8216, solves this with a master playlist that offers the same video as several variant streams at different bitrates. The player starts with a stream the connection can handle and switches as conditions change. Serve those segments from a CDN and the first seconds come from a server near the viewer.

What to look for: at least three renditions (for example 1080p, 720p and 480p), CDN delivery, a compressed poster that loads early, and a player script that doesn't block the page. To check it, open the page on a mid-range phone with network throttling turned on in your browser's developer tools, and time it from load to the first moving frame. Repeat on desktop. Those two numbers are your baseline.

RevvyScales transcodes every upload into a 1080p, 720p and 480p HLS ladder delivered from a CDN. Poster weight, scripts and tags are the rest of the story, covered in our guide to VSL page speed.

Muted autoplay, the unmute prompt and restart on unmute

The browser decides whether your video may start with sound, not you. Chrome's autoplay policy, in force since Chrome 66 in April 2018, always allows muted autoplay. It allows sound only if the visitor has interacted with your domain, has crossed Chrome's Media Engagement Index threshold on desktop, or has added your site to their home screen or installed it. WebKit's iOS policy lets a muted video autoplay, and on iPhone the video needs the playsinline attribute to play inside the page instead of going fullscreen.

So on cold traffic, design for a muted start in three steps:

A muted start in three steps

1. 1Start muted, instantlyMotion shows the visitor the video is already running, so the next decision is about sound, not about pressing play on a long video.
2. 2Ask for sound in plain wordsA large, high-contrast "Tap for sound" prompt over the video, with the presenter's face visible behind it.
3. 3Restart at 0:00 on unmuteA visitor who taps at second nine has missed the hook. Rewinding means they hear it from the first word.

Once the video autoplays, play rate stops telling you much, because almost everyone "plays". Watch unmute rate instead: visitors who unmute divided by unique visitors.

Say you split 4,000 visitors between your current prompt and a bigger one centered on the presenter. With the old prompt, 1,100 of 2,000 unmute (55%); with the new one, 1,260 of 2,000 do (63%). That's 160 more people hearing your hook from the same spend.

RevvyScales handles the mechanics on every plan: smart muted autoplay starts the video as a muted teaser, a first-time viewer who taps for sound is rewound to 0:00, and returning viewers resume from the deepest point they watched (player conversion features). Browser rules, prompt design and fallbacks are in the VSL autoplay guide.

Controls: seek bar, progress bar and pause screen

Every visible control hands the viewer a decision. In a VSL, two controls shape the outcome more than the rest: the seek bar and the progress bar.

The seek bar. If viewers can scrub, some will jump to the price without the argument that justifies it, and some will drag to the end just to see how long it runs. Hiding the seek bar keeps the story in order. The cost is control: viewers can't rewind a point they missed, and some will feel boxed in.

Hide seeking on long VSLs for cold traffic if your tests support it, but never hide pause, volume or captions. WCAG success criterion 2.2.2 (Level A) asks for a way to pause moving content that starts automatically and lasts more than five seconds, and the W3C lists synchronized media among its examples (Understanding 2.2.2).

The progress bar. A standard bar shows the true share of the video that has played. On a 40-minute VSL it barely moves in the first minutes, which tells a cold viewer exactly how big the commitment is.

A non-linear bar fills faster early and slower later. It changes how long the video feels; it changes nothing about the offer. If you use one, use it honestly: never next to claims like "this short video", and state the real runtime on the page. The options, the research and the tests are in the VSL progress bar guide.

The pause screen. A frozen mid-blink frame is a poor last image. A pause screen that shows your poster, or a custom image with a reason to keep watching, gives the viewer something to come back to.

In RevvyScales, hiding the seek bar, muted autoplay and the smart progress bar (you set how long the fast phase lasts and how far it fills) are per-VSL settings (VSL settings), and pausing shows your pause image instead of a frozen frame, or a built-in pause screen when you haven't set one. Here is a VSL's settings page, with the settings this guide covers numbered:

1234

1. 1Player, Tracking and Security tabsThe Player tab holds the seek bar, autoplay and smart progress bar settings, Tracking holds the pixels and Security holds the domain lock.
2. 2Video protectionIt shows the state of the protected streaming version made from the upload; while it is being prepared, the video plays in standard quality.
3. 3Scheduled visibilitySet a time for the page to go live or to expire.
4. 4Custom URLServe the VSL at a custom path on your connected domain.

Timed CTA: reveal the offer at the pitch

A buy button that's visible from the first second invites people to skip to a price they haven't been sold on, and it pulls eyes off the video. That's the reason to hide the button and the offer until the pitch: the moment the video first presents the offer and its price.

Take the reveal second from the final edit, not the script draft, and check it against your retention graph. If only a small share of viewers reach it, the problem is the content before it, and moving the button earlier treats the symptom. If you test an earlier reveal anyway, judge it by revenue per visitor, not by clicks, because an early button can collect clicks from people who only want to see the price.

What appears with the button matters as much as when: the offer block with any bundles, the guarantee, and the terms that qualify the price, placed right beside it.

In RevvyScales the offers, guarantee and CTA stay hidden until the pitch second you set, and the retention chart marks that second, so you can see how many viewers are still there when the button appears. Picking the second, mobile placement and measurement are in the delayed CTA button guide.

Pause, resume and exit intent

Long videos get interrupted: a call, a notification, a closed tab. Three behaviors decide whether those viewers make it back to the pitch.

• Resume for returning viewers. Someone who watched 18 minutes yesterday shouldn't start at 0:00 today. Resume from the deepest point they reached. First-time viewers still start at the beginning.
• A pause screen with a reason. Show the poster or a pause image that hints at what's coming, not a frozen frame.
• One respectful exit prompt. Asking once, as a visitor moves to leave, whether they're sure gives you one last chance. Asking repeatedly, or making the page hard to leave, spends trust you need at the checkout.

RevvyScales resumes returning viewers from the deepest point they watched (using a viewer cookie), shows the pause poster, and can show an exit-intent gate. Follow-up by watch depth, through email or retargeting, is covered in the pause and resume guide.

Before the next features, a quick check on the sound, control and offer decisions so far:

Check yourself: sound, controls and the offer

1. A first-time viewer taps for sound at second nine. What should the player do?

   Keep playing from second nineRestart at 0:00Show the pause screen
   Show the answer

   Restart at 0:00. A visitor who taps at second nine has missed the hook. Rewinding means they hear it from the first word.

2. Your VSL autoplays muted. Which number tells you whether the start works?

   Play rateUnmute rateAverage watch time
   Show the answer

   Unmute rate. Once the video autoplays, play rate stops telling you much, because almost everyone "plays". Unmute rate is visitors who unmute divided by unique visitors.

3. You hide the seek bar on a long VSL. Which controls stay?

   Pause, volume and captionsNone, the video plays on its ownOnly the unmute prompt
   Show the answer

   Pause, volume and captions. WCAG success criterion 2.2.2 (Level A) asks for a way to pause moving content that starts automatically and lasts more than five seconds. Never hide pause, volume or captions.

4. Only a small share of viewers reach your pitch second. What does that point to?

   The button should appear earlierThe content before the pitch
   Show the answer

   The content before the pitch. If only a small share of viewers reach it, the problem is the content before it, and moving the button earlier treats the symptom.

5. A viewer watched 18 minutes yesterday and comes back today. Where should playback start?

   At 0:00, so they hear the hook againAt the deepest point they reached
   Show the answer

   At the deepest point they reached. Someone who watched 18 minutes yesterday shouldn't start at 0:00 today. First-time viewers still start at the beginning.

Captions for sound-off viewers

With muted autoplay, your opening seconds are watched without sound by design. Captions keep those viewers following, help people watching in a second language, and are an accessibility baseline: WCAG success criterion 1.2.2 (Level A) asks for captions on prerecorded audio in synchronized media (Understanding 1.2.2).

Player captions (a separate VTT or SRT file) can be switched on or off and offered in several languages. Burned-in captions are always visible but fixed in one language and one size. A practical mix is burned-in text for the muted opening and a player caption track for the rest.

RevvyScales accepts VTT or SRT caption files per language on every plan; you can set one track to be on by default, and a CC button appears in the player. Formatting rules and caption tests are in the VSL captions guide.

Mobile and portrait video

Check your device split before you design anything. If phones dominate, the player has to work upright and one-handed: an unmute target big enough for a thumb, captions readable at phone size, and the buy button directly under the video when it appears.

A 16:9 video on an upright phone fills a strip across the top, about a third of the screen height or less. A 9:16 cut fills the screen but needs its own edit, because framing, text placement and b-roll all change. If you make one, serve it to phones and keep landscape for desktops. In RevvyScales a portrait companion video (Growth plan and up) does exactly that: phones and tablets get the 9:16 cut and desktops get landscape, on the same URL with the same analytics.

Playback speed and speed tests

A VSL doesn't have to play at the speed it was recorded. At 1.1x, a 30-minute video runs about 27 minutes and the pitch arrives sooner, so more viewers may reach it. Browsers correct the pitch of the voice by default when the speed changes (MDN on preservesPitch), so 1.1x sounds like a brisk speaker, not a cartoon.

The risks are comprehension and trust. Dense explanations, viewers listening in a second language, music cues and spoken disclaimers can all suffer when rushed. That's why speed is a test, not a setting.

On the Growth plan and up, RevvyScales can run playback-speed variants inside one VSL (for example 1.0x against 1.1x), each shown by weight and sticky per viewer (speed A/B test). Test design and how to read the results are in the VSL playback speed guide.

Protection: encryption, domain lock and watermark

VSLs that work get copied. The threats differ, and so do the defenses:

• Downloading the file. If the page exposes an MP4 URL, anyone can save it. Encrypted HLS streaming avoids handing over a file: the HLS specification defines an AES-128 method that encrypts each media segment, and short-lived signed key URLs make the keys hard to reuse.
• Embedding on other sites. A domain lock makes the player refuse to load anywhere you haven't listed.
• Screen recording. Nothing stops it. A watermark (your logo or a compliance mark) marks every recorded copy as yours.

Be clear-eyed about the limit: protection raises the cost of copying; it doesn't make copying impossible. RevvyScales streams AES-128 encrypted HLS with short-lived signed keys and never exposes a downloadable MP4. It can lock each VSL to the domains you list, with your connected custom domains added automatically, and adds an image watermark on the Growth plan and up. The full threat model is in how to protect a VSL from piracy.

Tracking: pixels, server events and click IDs

A player that can't tell you which ad produced a sale is a player you can't scale. The tracking chain has four links, and a break anywhere breaks attribution:

The four links of the tracking chain

A break anywhere breaks attribution

1. 1ArrivalThe page records campaign, ad set and ad, usually from a URL parameter.
2. 2On-page eventsPlays, milestones, the pitch and CTA clicks reach your ad platforms through pixels.
3. 3Click to checkoutEach CTA click gets an ID that travels to the checkout.
4. 4SaleThe checkout's thank-you page or a store webhook reports the sale back with that ID, and a server-side event tells the ad platform.

Server-side events back up the browser pixel. Meta recommends running its Conversions API alongside the pixel, and it deduplicates the pair when both carry the same event name and event ID, within 48 hours of the first event (Meta's deduplication docs).

In RevvyScales you set Meta, Google and TikTok pixels once on a folder or VSL, and the rvs URL parameter carries campaign, ad set and ad. Each CTA click gets a click ID, and Shopify, Whop, CartPanda or any checkout that can report that ID from its thank-you page closes the loop. On the Growth plan and up, Meta's Conversions API sends sales server-side, deduplicated so each reaches Meta once, and milestone webhooks (watched 25, 50 or 75%, pitch reached, completed, CTA click, lead captured) pass watch depth to your CRM or backend. The full setup is in the VSL tracking guide.

Analytics that tie watch time to revenue

Views and average watch time don't pay for ads. These reports change decisions:

• A second-by-second retention chart with the pitch marked, so you see where viewers leave and how many remain when the offer appears.
• Pitch reach: visitors who reached the pitch second ÷ unique visitors. It equals play rate × pitch retention (viewers at the pitch ÷ viewers who started).
• CTA click rate (clicks ÷ visitors who reached the pitch) and EPC (revenue ÷ CTA clicks): whether the offer converts the people who see it.
• Revenue per visitor (RPV): revenue divided by unique visitors, the number that combines everything above.

A worked example of why RPV gets the final word: two cuts of the same VSL each get 20,000 visitors. Cut A holds 31% of viewers to the pitch and makes 260 sales at an $80 average order: $20,800, or $1.04 per visitor. Cut B holds only 27% to the pitch but makes 320 sales: $25,600, or $1.28 per visitor.

Revenue per visitor: two cuts of one VSL

Worked example: 20,000 visitors each, $80 average order

Cut A31% of viewers held to the pitch · 260 sales · $20,800$1.04

Cut B27% held to the pitch · 320 sales · $25,600$1.28

A has the prettier retention curve. B makes more money from the same traffic, and at this sample the gap in sales clears 95% confidence in a two-proportion test. Ship B, then find out why its viewers buy more.

RevvyScales reports play rate, pitch retention, CTA clicks, conversion rate, orders, RPV, EPC and RPM, with a second-by-second retention chart that marks the pitch. Heatmaps of engagement across the timeline are on Growth and up. For the reading workflow, see how to read a retention graph.

A/B testing inside the player

Most VSL tests happen in the player, so testing should be part of it rather than bolted on. What to look for:

• Sticky assignment. A returning visitor sees the same variant, or your numbers mix.
• Weighted splits. You can send 50/50, or 80/20 while a risky variant proves itself.
• Money metrics. Variants judged on revenue per visitor, not just plays or clicks.
• Discipline about winners. A loser is cut on evidence, not on day two.

In RevvyScales, folders split traffic between whole VSLs by weight, sticky per new visitor, on every plan (folders and A/B testing). On the Growth plan and up you can also test inside one VSL: headline, poster, offer (bundles, prices, offer headlines) and playback-speed variants, each shown by weight and sticky per viewer (offer split tests). Here is a folder splitting new visitors evenly between two VSLs:

12

1. 1RPVRevenue per visitor of each VSL: the money metric that judges the test.
2. 2Traffic %Each VSL's weight and its share of new visitors; a visitor stays on the VSL they first saw.

Also on Growth and up, smart traffic split moves a folder's traffic toward the VSLs earning the most per viewer, and bring-your-own player lets you split-test an outside player on the same page. Test design that holds up is in VSL A/B testing.

Custom domain and page control

Your VSL's address appears in the ad preview and the browser bar. A custom domain with automatic SSL keeps the brand consistent from ad to checkout, and it keeps your URLs yours if you ever change hosts.

Scheduling matters more than it looks: a page can go live or expire at set times, which suits launches and promotions that really end. RevvyScales connects custom domains with automatic SSL, can put a VSL, folder or presell at a custom path, and can schedule a page to go live or expire.

VSL player checklist: must-haves and nice-to-haves

Use this table to compare players, or to audit the one you have. "Must" means must for paid cold traffic to a long VSL.

How to test a VSL player in one afternoon

Don't trust a feature list, including this one. Run your real VSL through these checks before you send traffic:

Test a VSL player in one afternoon

• Open the page on an iPhone and an Android phone on mobile data. Time it from load to the first moving frame.
• Confirm the video starts muted and the unmute prompt is readable at a glance.
• Tap for sound at second eight. The video should restart at 0:00.
• Pause. You should see a poster or pause image, not a frozen frame.
• Watch past the pitch second. The button and offer should appear directly under the video on the phone screen.
• Close the tab at minute five and come back. Playback should resume near minute five.
• Open the browser's network panel. You should see small streaming segments, not one MP4 file.
• Embed the player on a domain you haven't listed. It should refuse to play.
• Click the CTA, place a test order, and confirm the sale shows up against that click and ad.
• Check your analytics the next day for the visit, the pitch event and the sale.

0 of 10 done

Anything that fails here will fail at scale, with your ad budget running.