## 1. Executive Summary

Marcelo Tavora is co-founder of VK Digital and VK Matrix, a Brazilian launch operator who went from a sales internship at a 20-person content agency to his first R$1 million on Hotmart in August 2018, two years after his first affiliate sale. His path: paid traffic for affiliate offers (Facebook, then a 180-campaign Google Ads machine), then a pivot to co-producing expert launches triggered by a single phone call from an expert whose launch he topped as an affiliate, netting a 5% revenue split on R$2.7M gross with only R$50K in traffic. He then built the launch stack he wished existed: VK Matrix, a lead-scoring and attribution tool that joins UTM data, persona surveys, and checkout records to expose which audience and creative actually produce buyers. The core alpha of this episode is three-layered: (1) migrate from affiliate (a revenue model) to co-producer (a business) once you can run traffic, (2) warm audiences with a long, "new-fact"-driven pre-launch calendar instead of discount pressure, and (3) score leads with negative correlations so you stop paying for traffic that can never convert. He closes with a decision framework cost, reversibility, worst case, probable upside he uses to override fear and self-sabotage.

## 2. Alpha / Sauce

- **Affiliate is a revenue model, not a business.** As an affiliate you don't own the game: commissions get cancelled, producers share your ad structures with rivals, and rules can change overnight. Moving next to the expert closer to the power core meant less risk and more money on the same work.
- **One cold call changed the career.** He topped an expert's launch as an affiliate; she called, they met in person, and he negotiated 5% of gross to run copy, pages, email, and affiliate management. Result: R$2.7M gross on R$50K traffic, roughly R$100K+ profit after tax for ~3 months of part-time work.
- **Warm-up = a calendar of "new facts."** The winning launch had a 4-month pre-launch: a 2-class mini-course in September, a 5-day multi-expert event in October, more live events in November, sale in December. Each event is a new reason to touch the audience and raise awareness not another discount message.
- **Weekly lives beat content calendars.** He runs a fixed Wednesday live (kept sacred he refuses bookings that day), monthly if weekly is impossible, with gamified attendance prizes. Topics serve the persona, not the product, and each live ends with a soft pitch (a pit) to a free in-person event.
- **The VK Matrix attribution loop.** Capture page takes name/email/WhatsApp thank-you page carries the email in the URL persona survey response is joined to UTM source later joined to checkout by email/WhatsApp. One launch hit a 94% survey response rate, so he could name the exact audience and creative behind each buyer.
- **Negative correlations are the real lead score.** Answers that reduce purchase probability are what filter leads. One at-home fitness launch: a static ad brought over 50% of leads who "prefer training at the gym" a negative correlation. They rewrote the ad copy mid-capture ("5 days of free training at home") and flipped the mix to 80% home-preferring, far better-qualified leads.
- **Slice traffic by temperature before scaling.** A nail-design launch burned R$130K on cold audience for R$32K back; after installing attribution they spent only R$13K on cold for the same R$33K, while hot audience returned 1:1. The next launch hit similar gross with R$100K less traffic. Unexamined cold spend is where half of marketing budget dies.
- **Start today with automation services, not a personal brand.** His answer for a beginner: sell WhatsApp/CRM automation to local businesses (clinics, restaurants) R$2-3K setup plus R$600-800/month maintenance because every operating business already has traffic to optimize. Alternative: build audience, the single biggest digital asset; it took his service agency from zero to R$100K MRR in 4 months in 2023.

## 3. Methods & Setup

1. Get positioned inside a small company first his internship at a 20-person agency that grew to 120 in 18 months taught more than the degree; watch founders work up close.
2. Pick a simple paid-traffic entry: one affiliate offer, Facebook Ads at night after the day job; 2 hours an evening was enough to reach R$50K/month gross within 5 months.
3. Layer a second channel where competition is thin: in 2016-17 almost no producers advertised product names on Google, so he bought cheap keyword clicks and ran up to 180 distinct campaigns, sometimes four ad accounts bidding on the same term to find the cheapest quality-index win.
4. Use the producer's affiliate support dedicated managers reviewed his accounts and shared pixels; learn on someone else's dime.
5. When traffic works, convert from affiliate to co-producer: take a revenue-share deal (5% of gross) and run the whole operation copy, landing pages, email, affiliate coordination.
6. Build the long warm-up: mini-course (2 classes) 3 months out, multi-expert 5-day event 2 months out, additional lives 1 month out, all with a pit to the next step; fund the events inside the traffic budget.
7. Keep one fixed weekly live slot on the calendar; choose topics by persona pain, not by product angle; close with a pitch to a free in-person event as the next rung.
8. Instrument every link: UTM parameters on all ads, everything written to a database; capture page collects name, email, WhatsApp.
9. On the thank-you page, pass the email inside the URL (he does it with Elementor) so the persona-survey response auto-merges with the lead record no manual matching.
10. Cross-reference survey + UTMs against checkout by email and WhatsApp to attribute each sale to audience and creative; accept that it is never 100%, but 94% is enough to act.
11. Score leads in four buckets (disqualified, low, medium, high) and hunt negative purchase correlations; rewrite creative mid-capture when the disqualified share spikes on a specific ad.
12. Hold a story bank (uploaded to ChatGPT) of personal stories where money was pivotal e.g., paying for his father's funeral for hooks and live sessions; first hook and title, then the content, never the reverse.
13. Before any scary decision, run the framework: cost (time, effort, money), reversibility, worst case, probable positive outcome; if the worst case is survivable and reversible, the fear is self-sabotage; setup: fixed Wednesday live, UTM pass-through on the thank-you page, four-bucket lead dashboard.

## 4. Tools & Stack

| Tool | Role in the workflow |
| --- | --- |
| VK Matrix | Lead scoring, negative-correlation analysis, UTM-to-checkout attribution for launches |
| Meta Ads (Facebook/Instagram) | Cold and warm traffic for launches and evergreen funnels |
| Google Ads | Product-name keyword arbitrage; up to 180 concurrent campaigns in the affiliate era |
| Elementor | Passes the lead email in the thank-you URL to auto-join survey answers to the record |
| Hotmart | Checkout, affiliate network, and survey-triggered post-purchase flows |
| ChatGPT | Hook brainstorming and storage of his personal story bank for lives and creatives |
| WhatsApp / chatbot automation | Pre-qualification of inbound leads; the service he recommends beginners sell |
| ManyChat (legacy) | Early affiliate-era Messenger broadcasts with affiliate links (historical) |

## 5. Actionable Checklist

- [ ] Decide explicitly whether you are building a revenue model (affiliate) or a business (co-producer/product) and re-decide every quarter.
- [ ] If you can run paid traffic, pitch one expert a revenue-share deal to run their next launch end to end.
- [ ] Build a 3-4 month pre-launch calendar: mini-course, multi-expert event, extra lives, then the sale.
- [ ] Give every live and event a "new fact" angle story, guest, announcement instead of repeating the offer.
- [ ] Lock one recurring weekly slot and refuse bookings that conflict with it.
- [ ] Tag every ad with UTMs and write sources into one database.
- [ ] Add a persona survey on the thank-you page and pass the email through the URL.
- [ ] Join survey answers to checkout records by email and WhatsApp; report the match rate.
- [ ] Build a four-bucket lead dashboard and list every negative purchase correlation per offer.
- [ ] When one ad's disqualified share spikes, rewrite its copy during capture, not after.
- [ ] Split budget by audience temperature; kill or cap cold segments that don't return.
- [ ] Before any fear-driven decision, write: cost, reversibility, worst case, probable upside.