## 1. Executive Summary

Leonardo Fraga's VSL sold 35% more after changing exactly one element: the autoplay image that appears at the beginning of the video, before anyone presses play. Everything else stayed in place same copy, same video, same traffic, same offer. The problem this test fixes is that a lot of people simply do not pay attention to that opening image and end up using the same little preview everyone else uses; when every creative in the market carries the identical frame, that image saturates and instead of increasing conversion, conversion actually drops. The fix is to treat the preview like any other testable asset. VTurb's autoplay test lets you go in and put several images up as autoplay candidates; it then identifies which versions sell more and which sell less, and once statistical significance is reached you know for sure which one is the winner. Traffic to the weak versions stops, and all of it goes to the version putting the most money in your pocket. That is exactly what Fraga did, and the lift came in above 35%. The clip translates the money plainly: R$100 thousand becomes R$135 thousand for the same offer. The effort required is practically zero go into VTurb, create the autoplay test, select the versions, go and this is only one of seven tests the clip says you can apply to a VSL, most of them simple enough to run the same day.

## 2. Alpha / Sauce - what actually moves the needle, and exactly how to run it

- **The autoplay image is a conversion lever that everyone leaves on default.** It is the first thing a visitor sees before deciding to press play, it sits at the exact moment the purchase decision begins, and it is the piece most operators never think to change.
- **Default images saturate the market.** When a lot of people use the same preview, that image saturates and, instead of increasing conversion, conversion actually drops so ignoring this element is an active loss, not a neutral default.
- **Test images instead of guessing.** The autoplay test in VTurb holds several images at once and reports which version sells more and which sells less, so the decision comes from revenue rather than from whoever has the strongest opinion about design.
- **Stop only at statistical significance.** The result is not decided by an early glimpse of the numbers; once statistical significance is reached, you know for sure that the version you are looking at is the real winner and not a temporary blip.
- **Kill the losers, concentrate on the winner.** You stop sending traffic to the bad versions and send traffic only to the best version the one that is putting the most money in your pocket which is where the lift actually comes from.
- **The result reads directly as money.** More than 35% on conversion is the difference between making R$100 thousand and making R$135 thousand, with no other change anywhere in the operation to explain it away.
- **The effort is practically zero.** Go into VTurb, put the autoplay test in, select the versions you want to test and go for it no redesign, no new copy, no new traffic source, no production cycle.
- **It stacks with the other six tests.** The clip points to a follow-up video covering six more VSL tests that are also extremely easy and that anyone can do, so this is the first move in a sequence rather than a one-off.

## 3. Methods & Setup - Step-by-Step

1. Put the sales video on VTurb, since the autoplay test is a VTurb function; setup: the VSL hosted on VTurb with traffic already pointing at it before the test is built.
2. Open the autoplay test for that video; setup: a test slot created and ready to hold more than one image for the same video.
3. Collect several candidate images instead of keeping the default preview that everyone else in the market is already using; setup: distinct first frames lined up, each one visibly different from the others.
4. Load the versions into the test as autoplay options; setup: every candidate registered against the same video so the comparison stays clean.
5. Send your normal traffic to the video; setup: unchanged copy, unchanged audience, unchanged budget only the first image the visitor sees is varying.
6. Let VTurb measure which versions sell more and which sell less; setup: conversion recorded per version, so the ranking is made on revenue rather than on clicks or on looks.
7. Wait for statistical significance before drawing any conclusion; setup: no early stopping the test is only valid once significance is reached and the winner is certain.
8. Stop traffic to the losing versions; setup: all remaining traffic routed to the single winning image so nothing is wasted on the weak frames.
9. Keep the winner permanently on the video; setup: the winning preview left in place as the default autoplay instead of rotating back to the old default.
10. Measure the money at identical spend: +35% turns R$100 thousand into R$135 thousand; setup: a before/after comparison covering the same period and the same budget on both sides.
11. Roll the same procedure across every other VSL you operate; setup: each sales video queued for its own autoplay test rather than sharing one winner across unrelated funnels.
12. Continue with the six other conversion tests referenced in the clip; setup: an ordered backlog so the lifts keep compounding after the preview is fixed.

## 4. Tools & Stack - what each one is for here

| Tool | What it is for here |
|---|---|
| VTurb | Hosts the VSL and holds the test inside the video itself |
| Autoplay test | Runs the image versions and reports which one sells the most |
| Autoplay images | The candidate previews competing to be the first frame |
| Statistical significance | The signal that tells you the winner is real and the test can stop |
| Winning version | The image that receives all traffic after the test closes |
| Your VSL | The asset under test copy, edit and offer all left untouched |
| Follow-up video (six more tests) | The remaining easy VSL experiments to run next |

## 5. Actionable Checklist

- [ ] Confirm your sales video is hosted on VTurb.
- [ ] Open the autoplay test for that video.
- [ ] Collect several candidate first-frame images do not keep the default everyone uses.
- [ ] Register each image as an autoplay version in the test.
- [ ] Keep copy, traffic and budget untouched so only the image varies.
- [ ] Run the test until it reaches statistical significance.
- [ ] Do not call the winner early wait for significance before deciding.
- [ ] Stop traffic to every losing version.
- [ ] Route all traffic to the single winning image.
- [ ] Compare revenue before and after at identical spend: R$100k to R$135k.
- [ ] Apply the same test to each of your other VSLs individually.
- [ ] Watch the linked video for the six remaining conversion tests and queue them.