Video Sales Letter: What a VSL Is, How It Is Built, and What the Edit Is Doing

A client signs, the funnel gets built, and someone says the page needs a VSL. Then the questions start. How long, talking head or slides, and who cuts it.

A video sales letter is a sales pitch on video with one job: get the right viewer to click one button. Three things decide whether it does. The first thirty seconds, the length matched to how cold the viewer is, and the page around the player. Everything else is polish.

If the format is new to you, that is the normal state. Designers ask whether a fifty minute sales video can make sense, marketers ask whether VSLs still work at all, and both get twenty answers. This guide is the version for an agency that has to ship one for a client this month.

A VSL is a sales pitch you press play on

A VSL is a written sales letter read to camera, or over slides, or over stock footage. It sits on its own page, and the only link on that page is the ask. For a lead-gen agency the ask is almost always a booked call.

The video does the selling before the call. That is the whole reason it exists. Agencies who stopped building plain landing pages for high-ticket offers did it because the call was turning into the pitch. Move the pitch into the video, and the call becomes a fit check.

It is a letter, not a commercial. Copywriters who compare a sales letter with a VSL land on the same answer. The words are the same words. The video just controls the pace, so the reader cannot skip to the price.

The funnel has three shapes. Some run ad to opt-in to VSL to call. Some skip the opt-in and send the click straight to the video. One founder who booked 116 qualified calls in thirty days ran both a VSL page and a freebie page side by side. Pick one shape and measure it for a month before you change it.

The first thirty seconds are meant to lose people

The opening of a VSL is not there to keep everyone. It is there to keep the right people and send the rest away before they book a call they will not show up for.

1. 1Steep fall inside 30 seconds: the wrong people leaving
2. 2Gentle slope after it: the right people staying to the ask
3. 3A small step at the ask: normal, this is where the pitch ends

Qualify hard, early. A media buyer helping a coaching client watched show-up rate go from 2% to 8% after the hook was rewritten to say who the offer was for. Nothing else in the funnel changed. Put the who-this-is-for line in the first thirty seconds, on screen as well as spoken.

The early drop is not a bug. Agencies that run VSL funnels for a living treat the fall in the first half minute as filtering. As one funnel builder put it, "Someone who bails in the first 30 seconds was never a real lead", and the thread behind that line is worth reading in full. Judge the curve after the thirty second mark, not before it.

Watch drop-off, not clicks. An advertiser who ran their first video ads with no experience found the drop-off before second three told them more than any click rate. The same is true on the VSL page. Open the watch stats in your video host before you open the ad account.

Length follows the traffic, not a rule

There is no right length. There is a right length for how much the viewer already trusts you when they press play.

1. Solid: the low end of the range people run
2. Tint: the high end of the range

Cold traffic gets the long one. An email marketer who runs VSLs at both ends of a funnel keeps the rule short: "Cold traffic = Longer VSL. Warm traffic = can be a shorter VSL." A stranger from an ad needs the whole letter. A lead who asked for the call needs the summary.

Two videos often beat one. A high-ticket coaching funnel runs a twenty minute VSL beside a short four minute one, and a direct-response copywriter tells DTC brands to drop in a short five or six minute VSL first to warm people up before the offer. For a B2B pipeline, a mini VSL under ninety seconds has replaced whole email sequences. Cut the long one, then cut a short one from it.

Length is the wrong argument. A marketer tired of the debate argued that everyone obsesses over length and misses what matters, and the length threads in the copywriting forum never settle it either. What matters is whether the video holds the right viewer to the ask. Measure that, and let length follow.

The shape is the same at every length

Every VSL that converts runs the same six parts in the same order. The parts stretch and shrink with the length. They never reorder.

Start with who it is for. A copywriter who builds VSL templates says the whole job is getting people to self-identify. The first question the video asks should make the right viewer think, that is me. Write that line before anything else.

Problem, then the new way, then proof. A founder who turned six million Reels views into a six-figure month describes the middle as reframing the problem, showing the real solution, then showing proof. A media buyer who collected $560,000 from one campaign calls the offer section the promise. Put the numbers, names and screenshots in the proof part, on screen, not just in the voice.

One ask, said twice. Say the call to action when the proof ends and again at the close. Do not add a second offer, a second link or a second button. A simple VSL template that gets passed around has the same six beats and nothing after the ask.

What the edit is doing in a VSL that works

The script decides what is said. The edit decides whether anyone keeps watching while it is said. When agencies brief an editor for a VSL, the brief is nearly always the same four things.

1. 1Proof on screen under the claim, with the timestamp it belongs to
2. 2Captions styled to brand, kept clear of the face
3. 3The speaker, framed for a phone
4. 4The ask on screen when it is spoken

Captions that never cover the face. A brief for two VSLs of eleven and seven minutes asks for "Kinetic/animated captions, styled to brand, never blocking the speaker's face." Most viewers on a phone start with the sound off, so the captions carry the hook. Style them once per client and reuse the style on every cut.

Proof under every claim. The same briefs ask for b-roll and stock layered under each claim and story. A brief to finish a half-done ten minute VSL lists add b-roll where needed and standardise fonts. When the speaker says how many calls a client booked, the screenshot is on screen. When there is nothing to show, light motion graphics only where they support the point is the rule.

Pace, then cut the dead air. A talking-head and VSL editor's own listing says it plainly: remove dead air, long pauses and repeated lines, with clean jump cuts that do not feel jarring. Across the VSLs we cut for agencies, the passage that loses viewers is almost never a wrong argument. It is a stretch where the camera stops moving and nothing appears on screen. Thirty seconds of that costs more than a weak line.

Voiceover plus stock is a real format. Not every VSL has a face. Editors get asked to build VSLs from a voiceover and stock footage, and a marketing team with a twenty-five minute VSL wanted kinetic text to break it up. It works, but it needs a cut every few seconds to survive. Budget more editing hours per minute for it, not fewer.

Raw beats polished in the ad, not on the page

The forums disagree about production quality, and the disagreement is useful once you notice the two sides are talking about two different videos.

In the ad, raw often wins. Media buyers report that Meta rewards ugly creatives right now, and that for a high-ticket trading offer plain testimonial angles beat polished ones. A B2B marketer's under-ninety-second VSL works, in their words, because "It's not polished, but it works because it feels like someone recorded it for you." You can read their whole approach in the thread.

On the page, clarity wins. The other side has evidence too. For a luxury client spending $400,000 a month, polished photography beat raw content for cold traffic. And a salesperson proposing a VSL to a client went looking for references where the editing quality is high, because that is what the client expects to see.

Our side of it. Raw belongs in the ad, where the job is to stop a thumb. By the time someone is on the VSL page they have clicked, and the job is to be believed for five or twenty minutes. There, clean captions, proof on screen and steady pacing are what keep a claim watchable. Shoot it simply and edit it properly.

The page around the video decides whether it plays

A VSL that nobody presses play on has a play rate problem, not a video problem. The page decides the play rate.

1. 1Headline repeats the ad
2. 2First frame: a face and the phrase
3. 3One play button, autoplay off
4. 4The ask directly under the video
5. 5All of it above the fold

Turn autoplay off. A marketer who asked how to set up a VSL on a landing page ended up with autoplay off and short copy under the player, on the thread's advice. Autoplay with the sound off burns the hook before anyone chose to watch. A play button someone clicks is a viewer who chose.

Treat the first frame like an ad. An info-product funnel that tested thumbnails as image ads first took the two winners, at 7.5% and 10% click rates, and put them on the VSL page as the first frame. Another advertiser put a phrase on the landing video beside the play button after a month of low plays. Ask the editor for the first frame as a still, and test it the same way you test an ad.

Keep the player light and the button close. A developer found Vimeo iframes were hurting the page's speed on a VSL page with many videos. Load a thumbnail and swap in the player on click. Put the booking button directly under the video, above the fold on a phone, which is where high-intent traffic performs best.

Every extra step is a drop. One advertiser saw a 0.14% conversion rate sending traffic to a VSL with a schedule form asking for name, phone and email. A media buyer explained why: making people download a toolkit, watch a VSL and book a call adds three places to leave. One team removed the opt-in form and sent people straight to the VSL for the same reason. Count the clicks between the ad and the calendar, and remove one.

VSL or lead form is a real fork. Agency owners ask which to run, and the honest answer is a test. If leads are dropping between click and form, try a shorter opt-in before the full VSL. If calls book and then nobody shows up, the video is missing, and a VSL page for warm leads is the usual fix.

Read the symptom. No-shows point at the video. Cheap leads that never book point at the form.

What a VSL edit costs and how to brief it

Editors price VSLs four ways, and the threads where they are hired show the numbers. None of the numbers include the script or the shoot.

The cheap end is a gamble. A $60 budget for a three-minute direct-response VSL and a $25 per one-minute ad rate are real listings. So is the editor whose fifteen minute VSL with heavy graphics ran late while the client begged for it. A missed VSL deadline is a missed launch. Price the risk, not just the minute.

Per minute suits short, per hour suits long. A 20 euro per minute listing is fair for a four minute cut and brutal for a twenty minute one. Editors asked how long a stock-footage VSL takes answer in hours, because that is how the work runs. At our $18 an hour agency rate a five to seven minute talking-head VSL fits inside the free first four hours, and edits under about ten hours come back within 48 hours. A retainer, like the $500 a month direct-response listing, only pays off when the client sends work every week.

Send the script with the footage. Copywriters are surprised when a client expects them to narrate the VSL. The editor needs the script for the same reason: it says where each proof clip goes. A brief that gets a clean first cut has:

• The script, with a note beside every claim saying what goes on screen.
• The footage or voiceover, plus every screenshot, result and logo the proof part needs.
• One caption style per client, set once and reused.
• The exact call-to-action words, and the moment they are spoken.
• The first frame you want as a still, so it can be tested as an ad.

What to do about each pattern

Things to stop doing

• Autoplay with the sound off. It spends the hook on people who did not choose to watch.
• A second offer, link or button after the ask.
• Judging the video by the drop in the first thirty seconds. That drop is the hook doing its job.
• Arguing about length before you have watch stats from one version.
• A form between the ad and the video. Every field is a place to leave.
• Pricing a twenty minute VSL by the minute.

A fifteen minute check before each VSL ships

• Play it on a phone with the sound off. Can you tell who it is for by thirty seconds?
• Is every number, name and result on screen when it is spoken?
• Is there a stretch longer than twenty seconds with nothing new on screen? Cut it or cover it.
• Is the ask spoken and shown twice, and nowhere else?
• Is the booking button directly under the player, above the fold?
• Is autoplay off, and does the first frame carry a face and the phrase from the ad?
• Change one thing per version. One, so the watch stats can tell you what it did.

The short version

• A VSL is a sales letter on video with one ask, and it does the selling before the call.
• The first thirty seconds should lose the wrong people. Judge the curve after that mark.
• Cold traffic gets the long version, warm leads get the short one. Cut both from one script.
• Hook, problem, the new way, proof, offer, ask. Same order at every length.
• The edit is captions off the face, proof under every claim, no dead air, and the ask on screen.
• Raw wins in the ad. On the page, clear and steady wins.
• Autoplay off, first frame tested like an ad, button under the video, one step fewer.
• Per minute for short cuts, per hour for long ones, and never the cheapest listing on a launch deadline.

Read next

• White Label Video Editing: What Agencies Actually Pay, for how brand-neutral delivery works when the VSL ships under your name.
• How to Outsource Video Editing in 2026, for what to hand over and where agencies find editors.
• How Much Does Video Editing Cost in 2026?, for the per-hour, per-video and monthly numbers behind the table above.

You bring the script and the footage. The part that does not fit around client work is the cut, the captions and the proof layer, and that is what we do for agencies. Your first VSL edit is free up to four editing hours.

Last reviewed 15 September 2026. Every length, rate and result in this guide is from a marketer's or editor's own post, linked where it is used. Vimerse figures are the ones on our agencies page.