VSL playback speed: how to test 1.1x without hurting comprehension or trust
Playing a VSL slightly faster than it was recorded shortens the trip to the offer. This guide covers when a faster VSL playback speed helps, when it hurts comprehension and trust, how to run a clean 1.0x versus 1.1x test judged by revenue per visitor, and when to re-edit instead.
On this page
- What VSL playback speed changes
- Why a faster default can lift pitch reach
- When faster playback hurts
- How to run a VSL playback speed test
- How to read your speed test results
- Should viewers be able to change the playback speed?
- Should you re-edit or speed up?
- A playback speed test plan you can start today
- Frequently asked questions
VSL playback speed is the default rate your video plays at. At 1.1x a VSL runs about 9% shorter, so the pitch arrives sooner and more viewers can reach it, as long as comprehension and trust hold. Treat speed as a test, not a setting: run 1.0x against one faster variant, keep the CTA reveal aligned to the offer, and judge by pitch reach and revenue per visitor.
What VSL playback speed changes
Speed scales everything on the timeline at once: the runtime, the pitch second, the presenter's words per minute and every pause between them. The math is simple: new time equals original time divided by speed. It's one of the quieter VSL player features, and one of the cheapest to test.
The speed math
New timeoriginal time ÷ speed
New paceoriginal words per minute × speed
At 1.1x, a 30-minute VSL runs 27:16 (about 9% shorter), a pitch at 20:00 arrives at 18:11, and 150 words per minute becomes 165.
| Speed | A 30-minute VSL runs | A pitch at 20:00 arrives at | 150 words per minute becomes |
|---|---|---|---|
| 1.0x | 30:00 | 20:00 | 150 |
| 1.05x | 28:34 | 19:03 | 158 |
| 1.1x | 27:16 | 18:11 | 165 |
| 1.15x | 26:05 | 17:23 | 173 |
| 1.2x | 25:00 | 16:40 | 180 |
| 1.25x | 24:00 | 16:00 | 188 |
The 150 words per minute is an assumption for illustration; spoken delivery commonly falls somewhere around 140 to 160. Measure your own presenter: count the words in a two-minute stretch of the script and divide by two.
What doesn't change is the voice. When the playback rate changes, browsers correct the audio pitch by default (MDN on playbackRate); the preservesPitch property defaults to true and has been available across browsers since December 2023 (MDN on preservesPitch). At 1.1x your presenter sounds like a brisk speaker, not a cartoon, and lip sync holds because sound and picture speed up together.
Why a faster default can lift pitch reach
Picture a retention curve that, after the opening drop, slopes down steadily. If viewers leave at a roughly constant rate per minute, fewer minutes between the hook and the pitch means more viewers arrive at the pitch.
A worked example: 1,000 viewers are still watching at the one-minute mark, and from there 4% of the remaining viewers leave every minute.
Viewers who reach a pitch at 20:00
Worked example: 1,000 viewers still watching at 1:00, then 4% of the remaining viewers leave every minute
Half a percentage point of extra drop-off per minute wipes out the whole gain. That's the bet in one sentence: speed helps when viewers leave because of time, and hurts when it makes them leave because of what they can no longer follow.
Your retention graph tells you which kind of drop-off you have:
- Time-driven drop-off looks like a smooth, steady slope with no sharp cliffs. Viewers are running out of patience, not interest. Speed is a fair lever.
- Content-driven drop-off shows cliffs at specific moments: a slow story, a confusing mechanism, a weak transition. Speed won't fix those; editing will.
Here are the two shapes side by side:
Time-driven vs content-driven drop-off
Illustrative curves: share of the viewers at 1:00 who are still watching
- Time-driven: a smooth, steady slope
- Content-driven: cliffs at specific moments
Minutes into the video
Show the numbers
| Minutes into the video | Time-driven: a smooth, steady slope | Content-driven: cliffs at specific moments |
|---|---|---|
| 1:00 | 100% | 100% |
| 2:00 | 96% | 99% |
| 3:00 | 92% | 98% |
| 4:00 | 88% | 97% |
| 5:00 | 85% | 96% |
| 6:00 | 82% | 78% |
| 7:00 | 78% | 77% |
| 8:00 | 75% | 76% |
| 9:00 | 72% | 75% |
| 10:00 | 69% | 74% |
| 11:00 | 66% | 73% |
| 12:00 | 64% | 72% |
| 13:00 | 61% | 53% |
| 14:00 | 59% | 52% |
| 15:00 | 56% | 51% |
| 16:00 | 54% | 50% |
| 17:00 | 52% | 49% |
| 18:00 | 50% | 48% |
| 19:00 | 48% | 47% |
| 20:00 | 46% | 46% |
How to read a retention graph walks through finding the cliffs.
When faster playback hurts
Speed compresses everything evenly, including the parts that need room. Watch for these:
- Dense explanations. A mechanism, a run of numbers or a technical term needs processing time. A line that was clear at 1.0x can blur at 1.15x.
- Viewers who aren't like your team. Don't assume your own comfort speed is your buyer's. If your buyers are older than your team, or listen in a second language, test smaller steps such as 1.05x and read results by segment where your data allows.
- Delivery and emotion. Pauses are part of persuasion: the beat before a reveal, the silence after the low point of a story. Speed shortens every one of them.
- Music and sound design. Music beds and cuts timed to the beat speed up with everything else. A tight edit at 1.0x can feel frantic at 1.2x.
- On-screen text and captions. Burned-in text stays on screen for less time. Caption tracks stay in sync, but each line shows for about 9% less time at 1.1x, so viewers have to read faster.
- Spoken disclaimers. If your VSL speaks a disclosure, such as typical results or subscription terms, speeding it up changes its cadence. The FTC's digital disclosure guidance says audio disclosures need "a volume and cadence sufficient for a reasonable consumer to hear and understand it" (.com Disclosures). Listen to those lines at test speed. This isn't legal advice.
- Trust. A presenter who sounds rushed can sound like they're pushing. If the faster version wins on revenue but refunds rise, the speed may be selling harder than the product delivers.
How to run a VSL playback speed test
Six steps, in order; open each one:
A clean playback speed test
1Tighten the edit first
What happens: Test speed on your best cut, not on one you're about to change.
Why it matters: Speed multiplies whatever you give it.
2Pick two arms
What happens: 1.0x against 1.1x answers the main question fastest. Add 1.15x or 1.2x later, once you know which direction pays.
Why it matters: Every extra arm splits your traffic further.
3Keep the CTA on the offer
What happens: The reveal has to move with the speed. A pitch at second 1,200 of the video arrives after about 1,091 seconds at 1.1x (1,200 divided by 1.1).
What goes wrong:
- A player that counts seconds since the page loaded won't move the reveal. One that keys the reveal to the video's own timeline handles this automatically.
4Split by visitor, sticky, on the same traffic
What happens: Each visitor gets one speed on every visit.
5Decide the metrics in advance
What happens: Primary: revenue per visitor. Secondary: pitch reach (visitors who reach the pitch ÷ unique visitors) and CTA click rate (clicks ÷ visitors who reached the pitch). Guardrails: refunds and complaints.
What goes wrong:
- Leave completion rate out of the decision: a shorter video gets finished more often almost by definition.
6Run full weeks, and don't stop at the first good day
Why it matters: Checking daily and stopping the moment a variant looks ahead inflates false winners.
How much traffic does a speed test need?
Less for pitch reach than for sales. A worked example with the standard two-proportion sample-size formula, at 95% confidence and 80% power:
- To detect pitch reach moving from 25% to 28%, a 12% relative lift, you need about 3,400 visitors per variant.
- To detect purchases moving from 2.0% to 2.24% of visitors, the same 12% relative lift, you need about 56,600 visitors per variant.
That's roughly 17 times the traffic for the number that pays, which is why a speed test can show a clear pitch-reach answer long before a clear revenue answer. Use pitch reach to spot a loser early, and wait for revenue before you crown a winner. More on test design in VSL A/B testing.
Put in your own conversion rate and the smallest lift you care about. The calculator starts at the purchase example above:
Calculator
A/B test sample size calculator
How many visitors each version needs before a difference in conversion rate means something.
- Visitors in all
- Days at your traffic
- Orders per version at the current rate
95% confidence and 80% power, the usual defaults for comparing two conversion rates. Fix the sample before you start and read the result when it is reached: stopping the moment one version pulls ahead finds winners that aren’t there. With more than two versions, each is compared with the control and needs this many visitors.
On the Growth plan and up, RevvyScales runs playback-speed variants inside one VSL, for example 1.0x against 1.1x: each variant is shown by weight, sticks to the viewer on return visits and gets its own breakdown in analytics (speed A/B test).
How to read your speed test results
Answer two or three questions about your result to see what it likely means and the next step:
What your speed test result likely means
Pitch reach first, then revenue per visitor, then refunds
With the faster speed, what happened to pitch reach?
And revenue per visitor (RPV)?
Did refunds rise too?
Best fit
Drop-off was time-driven
Pitch reach up and RPV up, without refunds rising. Next step: keep the faster speed; consider testing one step higher.
Best fit
Selling harder than the product delivers
RPV up, refunds up. Next step: treat it as a loss until refunds settle.
Best fit
More viewers reach the offer, but a smaller share are convinced
Pitch reach up, RPV flat. Next step: check CTA click rate and EPC; try a smaller step.
Best fit
Speed isn't your lever
Pitch reach flat, RPV flat. Next step: go back to the edit and the retention cliffs.
Best fit
The faster pace is losing people
Pitch reach down. Next step: revert, then watch the opening at test speed.
Read segments with care. If the faster version wins on desktop and loses on mobile, that's a hypothesis for the next test, not a conclusion, unless each segment has enough data on its own. For the revenue formulas behind these readings, see revenue per visitor.
Should viewers be able to change the playback speed?
Viewers are used to choosing their own speed on platforms like YouTube (YouTube Help). A speed menu in a VSL has real trade-offs:
- For it: impatient viewers can reach the offer on their own terms instead of leaving, and slower listeners can slow down.
- Against it: with the seek bar hidden, a 2x option is a skip button in disguise, and every extra control is one more thing to tap besides the video.
If you offer one, key the CTA reveal to the video's timeline, so the button still appears exactly at the offer at any speed. Either way, the default speed is the one every viewer starts at, which makes it the lever worth testing.
Should you re-edit or speed up?
Speed and editing both shorten a VSL, but they cut different things.
Speed up or re-edit
Speed up
- What it cuts: every second, evenly
- Cost and time: a setting change
- Effect on delivery: shorter pauses everywhere
- Best for: a smooth, steady decline
- Reversible: instantly
Re-edit
- What it cuts: only the weak parts
- Cost and time: editing hours, sometimes a re-record
- Effect on delivery: strong moments keep their natural pace
- Best for: cliffs at specific moments
- Reversible: only if you keep the old cut
A decision rule: if the retention graph shows cliffs, edit those moments first. If it shows a smooth slope and the edit is already tight, test speed. The strongest setup is often both, in that order: a tighter cut, then a speed test on it. For cutting for retention, see VSL editing.
Speed isn't the only lever on how long a video feels; the progress bar is the other. See the VSL progress bar guide.
A playback speed test plan you can start today
Tick each step as you go; the ticks stay in this browser:
A playback speed test plan
0 of 7 done
Frequently asked questions
Does speeding up a video change the voice?
Not the pitch, in modern browsers. When a web video plays faster, the browser corrects the audio pitch by default, so the presenter sounds like a quicker speaker rather than a cartoon. What does change is the pace: words per minute rise, pauses shrink, and music and cuts move faster. At small steps like 1.1x the change is subtle; at bigger steps it becomes easier to hear.
Is 1.25x too fast for a VSL?
It can be, depending on how fast your presenter already talks and how dense the script is. At 1.25x, a presenter recorded at 150 words per minute delivers about 188, and every pause shrinks by a fifth. Rather than jumping there, test 1.0x against 1.1x first, then step up only if each step improves revenue per visitor without raising refunds.
Should I re-record the VSL instead of speeding it up?
Re-record when the problem is the delivery itself: a presenter who drags, long pauses everywhere, or filler that editing can't remove. Speed it up when the delivery is good but the whole video feels a little slow. Re-recording costs more but gives you a natural pace; speed is free and instantly reversible. A sensible order is to test speed first, because it's cheap, and re-record if the faster version wins clearly.
Does faster playback affect spoken disclaimers?
Yes. Anything spoken in the video speeds up, including typical-results statements, subscription terms and other disclosures. FTC guidance on digital disclosures says audio disclosures need a volume and cadence a reasonable consumer can hear and understand. Listen to those lines at the test speed, and if they sound rushed, re-record them more slowly or also show them clearly on screen. This isn't legal advice.
How long should a playback speed test run?
Until you have enough data, and always in full weeks. A shift in pitch reach can show up within days at moderate traffic, but revenue per visitor needs many times more visitors, because purchases are rarer than pitch views. Decide your sample size before you start, don't stop the first day one version looks better, and check refunds before you make the change permanent.
Can I speed up only part of a VSL?
Yes, in the edit. You can speed up slow stretches, such as b-roll sequences or a long backstory, while leaving the hook, the mechanism, the offer and any disclosures at natural pace. That takes more work than a player setting but avoids rushing the parts that need room. Keep the change subtle so the presenter's voice and movement still look natural.