## 1. Executive Summary

Gabriel Fogacao came out of black-hat forums and e-commerce competitive research, then built Brazil's first organized espionage method: 250+ one-on-one mentorships, a recorded course with 1,000+ students, validated by students who went from R$10k/month to investing R$160k/month in traffic. His thesis: espionage is not opening Facebook Ads Library and finding one offer it is reading the market at macro and demand level (he maps 1.3M Alzheimer patients in Brazil to the 4 5M relatives who would actually buy an activity e-book, then checks whether the US already sells it), and it is a source of study, not a source of Ctrl+C. The alpha of this episode is a fully free level-zero stack DeepTube transcription search, Google Ad Transparency, SimilarWeb, spreadsheet view-tracking that lets you watch a competitor burn R$10 20k/day on creative tests so you never pay that tuition yourself, plus the Google dorks that surface any operator's CNPJ and hidden pages, and the protection side: decoy offers, cloakers, and why the big players who never complain about plagiarism run a fake front while scaling behind it.

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

- **Level zero (all free): DeepTube Ad Transparency SimilarWeb spreadsheet.** DeepTube searches YouTube ads by transcript (type three words the ad says, not a paragraph more words, dumber search); take the winning ad's URL to Google Ad Transparency, where you watch every ad of that advertiser in real time; take the domain to SimilarWeb for monthly traffic (previous-month estimates, always). Then log it: day 1 = 50k views, day 2 = 200k means he started scaling you just read his test results live and saved his R$10 20k/day testing budget.
- **Ads Library secrets most people never use.** It searches the text written inside the image, the video description/transcription (keep it to ~3 words), and the display link not the destination URL. Put `globo.com` in quotes and you find anxiety ads routing to a VSL. Two consequences: you find hidden offers, and you can hide yours (see protection below). Search ad-libraries by the media buyer's CNPJ: `"<name>" CNPJ` in Google (Google hacking/dorks), then test each CNPJ in the library.
- **Google dorks + Yandex = the indexed underbelly.** Exact-match quotes; `site:domain.com` lists every indexed slug of a site (bypass WordPress `?sitemap` and `?s=` blocks); the disclaimer text of a Facebook campaign run in quotes surfaces everyone buying direct traffic on that template; Yandex (Russian Google) keeps indexed pages Google has dropped one page on Google, ten slugs on Yandex.
- **Spy platforms are mostly Semrush in disguise.** Mainstream spy tools query `pay.hotmart.com` backlink data, so anything with a redirect or unindexed checkout vanishes; free tools beat paid ones in several spots. BigSpy was once great for YouTube, now ~R$14k/year; a ~R$500/month technology-search tool (free tier = first 100 rows) lists every site running VTurb, Hotmart, ClickMagick or GTM so you can intersect them.
- **Level 1 2: walk the funnel, then unmask the operator.** From the creative, walk the whole funnel landing, upsell, downsell (find via sitemap or `?sitemap` on WordPress), other VSLs because the difference between what sells and what sells a lot is the backend: an offer converting on upsell transplanted into your operation lifts margin directly. Then find all their domains: same IP via viDNS dumps, shared Facebook pixel, ClickMagick code, or GTM tags he pulled 70 domains of one low-profile Nutra operator, one had a.br registration exposing the owner, and Instagram confirmed eight figures.
- **Model the language and psychology, never the story.** He read women in a Facebook group calling it "my rich little money" a phrase he'd never heard and put it in his VSL: "for women who tried everything and only lost their rich little money." Copy the market's words (the niche says tight clothes, not clinical terms), the root-cause argument, "the industry will take your money" every health VSL runs it because it works but change Juliana's trip to Europe. His proof: modeling one Hormozi proof-angle cut his cost per lead by 70%. Swipe files (swiped.co, hard-to-find-ads, swiped.com.br) are the archive of what already worked a habit that started with 1950s copywriters, not Brazilians.
- **Free market research before any tool:** open a viral expert video in your niche on YouTube and read the comments the persona writes her pain there ("tried Thai tea, tried encapsulation."), and so does every Facebook group. That is the copy research. Also mine non-English markets: a keyword in German finds scaled advertorials pulling 500k visits; "prostate" is spelled the same in German FB Aff Bank (free native spy) found him scaled German VSLs by accident.
- **Protection is a second-order problem sell first.** His rule to students: 100% of energy on selling beats 50/50 split with hiding. Basics only: block the sitemap, don't scale on WordPress (R$50 of hand-written HTML converts better and never crashes with your R$200k/day budget), point DNS through Cloudflare (free), isolate pixels/tags. When you're big: the cloaker parameter never appears in the library it only lands in the URL when the ad serves to a real user so save it when you see the ad and the cloaker is broken; a persona account (Chile VPN, 50+ woman interested in menopause, joining FB groups and Googling remedies until the algorithm serves scaled ads) is the 8020 for forcing the serve. White Rabbit (from ~$500) he never broke.
- **The decoy beats the wall.** The sophisticated players who never get plagiarized don't hide harder they show a fake front: a public page running an old VSL that no longer converts (or a retired creative as display image), so the plagiarist clones the loser while the real VSL scales on VTurb hosting where nobody can even view it. Same logic as a cloaker whose white page looks like a normal offer instead of a "not found" shrek image.

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

1. Define the market first: which platform the lead's first touch lives on (Facebook, YouTube, native blog) that's where you research; setup: one niche, one platform, written down.
2. Run the level-zero loop: DeepTube transcript search (3 words) grab URL Google Ad Transparency grab domain SimilarWeb traffic; setup: free accounts only.
3. Spreadsheet the ads: date found, view count, format, hook; revisit in 48h views doubling = scaling, falling = dying; setup: one sheet per niche, updated weekly.
4. Mine the language: comments on viral expert YouTube videos + Facebook groups; pull exact phrases, failed solutions, fears; setup: swipe doc of verbatim quotes before writing any copy.
5. Dork for operators: `"<name>" CNPJ`, `site:domain.com`, quoted disclaimer text; test each CNPJ/name in Ads Library; setup: keep a CNPJ list per player.
6. Do the demand check: search the problem's population, multiply by relatives/caregivers, then confirm an offer already exists (US first, then Germany/EU, then Brazil); setup: one demand hypothesis per idea.
7. Walk the competitor's funnel as the lead: creative landing upsell/downsell via sitemap second VSL; setup: note which backend pieces you could transplant.
8. Unmask the operator's full footprint: same-IP (viDNS), shared pixel, ClickMagick, GTM tag search; setup: domain dossier per player, refreshed monthly.
9. Log traction by month-over-month traffic and referrers in Semrush referrers reveal the affiliate army pushing one URL.
10. Build the swipe file: Drive/OneDrive, folders by niche and by format (VSL / creative / headline), download everything worth keeping.
11. Only after revenue: protection pass sitemap off, HTML pages, Cloudflare, isolated pixels, decoy front page with a retired VSL, VSL hosted where it can't be viewed.
12. If you run black/gray niches: persona account + residential proxy to force the ad serve and capture the cloaker parameter; setup: account aged as the exact avatar.
13. Weekly cadence: one hour of spying before spending; model one angle per cycle and measure CPL against control; setup: change one variable at a time.
14. Guard against information obesity: pick findings you will test this week, then close the tabs tools don't make money, applied patterns do.

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

| Tool | What it does here |
|---|---|
| DeepTube | Finds YouTube ads by transcript search; entry point of the level-zero loop |
| Facebook Ads Library | Keyword, in-image text, transcription and display-link search; free real-time creative feed |
| Google Ad Transparency + dorks | All ads of one advertiser in real time; `site:`, quotes and CNPJ lookups expose operators |
| SimilarWeb / Semrush | Traffic estimates, monthly trends, referrers (affiliate armies) per domain |
| viDNS / BuiltWith | Same-IP and technology fingerprints pixel, GTM, cloaker and stack discovery |
| Cloudflare | Free DNS masking; the first wall against same-IP lookups |
| White Rabbit | Hardest cloaker he tested; parameter only appears when the ad serves to a real user |
| VTurb | VSL hosting that keeps the real sales video unviewable to scrapers and plagiarists |

## 5. Actionable Checklist

- [ ] Run the DeepTube Ad Transparency SimilarWeb loop on three competitors this week
- [ ] Start a view-tracking spreadsheet and revisit each ad in 48 hours
- [ ] Read 50 YouTube comments in your niche and save verbatim phrases in a swipe doc
- [ ] Search your own Ads Library results by in-image text and by 3-word transcription
- [ ] Dork `"<competitor>" CNPJ` and test each CNPJ in the transparency libraries
- [ ] Check whether the US (and German/EU) already sells your offer before building it
- [ ] Walk one competitor funnel to the upsell via `site:` and sitemap queries
- [ ] Pull one operator's full domain list via shared pixel / GTM / viDNS
- [ ] Model one proven angle into your copy change the story, keep the psychology
- [ ] Archive every winner you download into a niche+format swipe file
- [ ] Before scaling, move off WordPress to HTML, mask DNS with Cloudflare, block the sitemap
- [ ] Ship a decoy front page with a retired VSL so plagiarists clone the losing offer