## 1. Executive Summary

Alvaro Ezequiel, 26, from a city of 27,000 in Sergipe, is one of the first Brazilian producers of global nutra by his factories' account, the first Brazilian to scale the category to eight figures; the episode intro puts his cumulative sales above 200 million. He has been in the market since 2019 but went all-in on digital around the end of 2022, and today a six-person company bills more than 30 million in a single month at a 59% profit margin, with 39% as its worst month ever. His doctrine is the boring basics done extremely well: ad to VSL to checkout, no tracker, no cloaker, every word of copy written by AI, and more volume of what already works. He claims more than $25 million earned with no copywriter on the team.

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

- **ABO to start, CBO to scale.** ABO is the best campaign format for a beginner on a low budget: R$500 of ABO spend made R$1,000-R$2,000 a day for him, and he tells beginners to start at R$500-R$1,000 a day. Above R$5,000 a day, move to CBO. He entered as an affiliate with R$10,000 first month: R$10,000 spent on ads, 7,000 returned, which the host tallied at about 35,000 billed and 3.5 ROI with the dollar at 5.15 and he says a producer seat needs at least R$100,000.
- **Raise budgets with a 5 AM rule, never by hand mid-day.** Launch at midnight on a low budget, then arm a rule: when the campaign spends X or makes X sales, set its budget at 5 AM to R$5,000-R$10,000. If the rule trips at 2 AM, Facebook starts recalculating the audience immediately, whereas a manual mid-day raise sits in a 2-3 hour recalibration window he ran one day with rules and one day without, and rules always produced higher ROI; a friend went from 1.6 to 2 by adopting them. For an afternoon raise, arm the rule two hours ahead. He has not raised budgets manually mid-day for almost two years.
- **Hybrid scaling: campaign volume one day, budget the next.** One day he adds campaigns, the next day he raises the budgets of those campaigns. This cadence is what protects the margin: best month 59%, worst month 39%, normal months above 40%.
- **Flip the winner's face and rerun it in the dead dayparts.** He found daytime slots with no sales, mirrored the character's face in his best-performing ad, kept everything else identical, and filled those gaps. That creative family sold more than $7 million; the original was a two-person dialogue, deliberately unlike the standard hook-body-CTA ad everyone runs.
- **The creative qualifies the buyer, it does not just tease.** Content-style ads that do not look like ads (podcast and news formats) work best for him, because curiosity-only traffic converts worse into the VSL; he models organic content mostly from YouTube and reads video comments for objections. Throughput: 150-250 creatives per week per traffic source, cutting 20-50 variations from each validated ad and expecting 3-5 to scale.
- **All copy comes from trained AI agents.** Two custom GPT agents one for VSL copy, one for creative copy trained on the material of established copywriters and on patterns extracted from proven VSLs. When he put a hired copywriter's ad against the AI's ad, the AI brought the better result, so he never hired one; that is how he earned more than $25 million with no copywriter on payroll. He feeds the agent winning leads, bodies, mechanisms and closes, then asks for a new VSL on top of what already worked.
- **Frankenstein your VSLs out of proven parts.** Run VSLs side by side, keep the best lead from one, the best body from another, the best mechanism and the best close from the rest, and stitch them he ran 21 VSLs simultaneously to do this. Today output is about 4 VSLs a month, split 50/50 between AI-built and Frankenstein, and 99% of his successful VSLs came from this process. It is also his bridge from affiliate to producer: run the top five VSLs in a niche, see which parts convert, then build your own funnel on that evidence.
- **Judge at account level and keep ROY above 2.** ROY is his metric after product, staff and tax costs (not ROAS), and acceptable is always above 2. If the account clears that number he touches nothing: campaigns that look dead in the morning sell at night, so with 40 campaigns in one account and only one selling, he still does not move pausing the morning losers would have cost him the rest of the day.

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

1. **Pick your geo matrix and launch all four setups on every account.**; setup: five countries United States, Canada, United Kingdom, New Zealand, Australia; the four setups are (1) US only, (2) US + Canada + UK (their mixed time zones keep the auction warm and give a lower CPM), (3) New Zealand + Australia, (4) all five together; upload 3-4 copies of each per account, roughly 12 campaigns in total.
2. **Validate creatives in ABO first.**; setup: one campaign, 10 ad sets, 3 ads per ad set, all countries targeted; the CBO version of the same test is one campaign, 5 ad sets, 2-3 ads each.
3. **Score the test on ROY inside 48 hours.**; setup: within a 48-hour window you find your champions; anything above ROY 2 is validated duplicate the winners, kill the losers, roll the structure into the next day.
4. **Pre-scale winners with micro-variations, same structure, same budget.**; setup: change exactly one element per variation face direction, hook wording, subtitle color, background, actor, call-to-action placement, aspect ratio (square has been his best performer).
5. **Scale by alternating volume and budget day by day.**; setup: day one, duplicate yesterday's winners to raise campaign count (12 to 15 means duplicating the 8 that worked, killing the 4 that did not); day two, raise budgets of those campaigns only through rules.
6. **Set the budget rule before you sleep.**; setup: launch at midnight with a deliberately low budget; rule reads "once this campaign spends X or makes X sales, set its budget at 5 AM to R$5,000-R$10,000"; a rule that trips at 2 AM makes Facebook recalculate the audience right away instead of waiting for 5 AM.
7. **Make every decision at account level and keep your stomach.**; setup: if the account is above ROY 2, do not touch it the same 10 campaigns can sell four in the morning, four in the afternoon and four at night; one account runs 20 campaigns at R$1,000 each; he has pushed R$20,000-R$30,000 into a day on some accounts and R$200,000-R$300,000 on others.
8. **Build the creative factory.**; setup: 150-250 creatives per week per traffic source; take a validated ad and cut 20-50 variations from it, expecting 3-5 to scale; the format that wins for him is square, with the first 5 seconds cinematic and everything after that typography black background, yellow letters on both Facebook and YouTube.
9. **Mine spy tools, organic feeds, comments and marketplace rankings.**; setup: Spy Hero and Deeptube, where he searches by the text inside an ad; an ad still running after about a month is probably profitable, and the ad that sold over $7 million had only 2 active placements and was 2 days old when he found it; he also models organic YouTube content, reads video comments for objections, and watches offers climb from page 5 to page 4 to page 3 of the marketplace as a live signal to test them.
10. **Ship VSLs on two tracks, 50/50.**; setup: track one is AI-built lead, body, mechanism and close prompted one piece at a time and assembled by hand; track two is Frankenstein best parts stitched from different winning VSLs; about 4 VSLs ship per month, two per track, and when he validates a new niche he runs up to 21 VSLs at once.
11. **Keep the funnel bare.**; setup: ad, VSL, checkout no tracker, no cloaker, because every extra layer makes the next outage harder to diagnose; offers run through Digistore, which sends 90% of events correctly, and offer managers still walk him through the winning VSL and ad angles over Skype.
12. **Buy YouTube traffic on Shorts.**; setup: vertical, under 1 minute so it runs as a Short instead of a pre-roll; spy tools do not index Shorts, so nobody can find your ad; 99% of his YouTube traffic is Shorts; the current channel split is about 90% YouTube to 10% Facebook; the first test he cites: R$800 spent, R$4,000 sold.

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

| Tool | Role | In this episode |
| --- | --- | --- |
| Facebook Ads | Main paid channel historically | ABO validation, CBO scaling, four geo setups, and the 5 AM budget rules |
| YouTube Ads (Shorts) | Now the dominant channel, roughly 90/10 over Facebook | Shorts only: vertical, under 1 minute, invisible to spy tools; 99% of his YouTube traffic |
| Digistore | Offer platform and network access | Sends 90% of events correctly; managers hand over the VSL and the working ad angles on Skype |
| ClickBank & ByGood | Public marketplaces | Where offers show up 2-3 months after they appear in closed, invitation-only networks |
| Spy Hero & Deeptube | Creative research | Deeptube searches ads by the text inside them; he lifts hooks and angles from long-running ads |
| ChatGPT + custom GPT agents | All copywriting: one agent for VSLs, one for creatives | Trained on copywriter material and proven VSL patterns; beat his hired copywriter; wrote the copy behind $25M+ |
| Poe | Multi-model AI marketplace | Secondary model access, though he keeps most work inside ChatGPT because it is cheaper |
| Flick (name as he recalls it) | Cinematography generation | He pastes the copy in and it generates the cinematic first-5-second shots for his ads |

## 5. Actionable Checklist

- [ ] Launch your first campaigns as ABO at R$500-R$1,000 a day; switch to CBO once you pass R$5,000 a day.
- [ ] Upload all four geo setups on every new account: US; US + Canada + UK; New Zealand + Australia; all five together.
- [ ] Validate every creative within 48 hours and keep only what is above ROY 2 (return after product, staff and tax costs).
- [ ] Move every budget raise into a rule that fires at 5 AM to R$5,000-R$10,000; never raise budgets by hand mid-day.
- [ ] Alternate your scale days: campaign volume one day, budget increases the next.
- [ ] Cut 20-50 variations of each validated ad and keep 150-250 creatives a week moving per traffic source.
- [ ] Mirror the face and restyle the first 5 seconds of your best ad, then rerun it in your dayparts with no sales.
- [ ] Train a custom GPT on copy you respect and make it write your ads and VSLs before you pay a copywriter.
- [ ] Keep the funnel bare: ad, VSL, checkout no tracker, no cloaker.
- [ ] Decide at account level only, and never pause campaigns while the account is above ROY 2.