Loading page…
$45K MRR in 60 days — validated with 200 landing pages and $5 deposits before writing any code
profitable from day one
"A waitlist signup shows interest; a credit card shows commitment" — while the product didn't exist yet, 200 landing pages + a $5 skip-the-line deposit revealed who would actually pay.
Fewer bars = easier, cheaper, or faster for an AI-assisted solo builder. Editorial judgments based on the case details.
[Video] Giovanni Brees, New York, zero development experience but a heavyweight background: ran a product-launch agency for over a decade, helped bring 4,500+ products to market, with over $500M in cumulative crowdfunding + direct sales — he walked in with a mature paid-acquisition methodology. The starting point was his own pain: 100+ emails a day. The product: AI drafts replies inside your inbox in your own voice, based on the email thread and uploaded company materials (price lists, FAQs). $45,000 MRR 60 days after launch, profitable from day one, with all revenue reinvested into ads. Now 11 people (6 devs + 3 marketing + 2 support, with AI handling most of support).
The customers existed before the product did. Brees built one landing page, then had Lovable spin out variations until there were more than 200 of them, mapped to roughly 15 main ICPs and their subgroups — a solo lawyer, a lawyer inside a firm and a lawyer working insurance each got their own visuals and copy. The logo took five minutes in PowerPoint: company name, two leaves on top, screenshot, done. Traffic came from Meta on a starting budget under $5,000, targeted by hand on interest, age group and location rather than handed to automated targeting, with creative built to interrupt a scroll — a solid colour background, one very short call to action, an emoji. Cost per lead was the number that mattered, and the retargeting pixel fired only on visitors who actually read the page, not on everyone who bounced after two seconds. Google Ads was deliberately kept out until launch: search intent is real, but someone searching wants the tool now, and a waitlist is a poor landing for that intent at a higher cost. Budget went up only on the pages that converted.
The first money arrived before a single line of product code. Joining the waitlist put you in a queue that started at number 1,000 — his own signup was 1,001 — and $5 moved you into the first hundred to get access at launch. That was the whole point: a signup is someone being polite, a credit card is commitment, and the prepayment rate per ICP, not the click rate, decided which audiences got built for. Referrals stacked on top, with five signups buying two extra months free. Where payment rates lagged, the fix was trust rather than discount — European segments, more wary about entering card details than American ones, got a follow-up email about how he used the assistant inside his own company, and their rates climbed back to match the others. Those $5 deposits were small individually, but together they repaid part of the ad spend before launch, which is why the business was profitable from the moment it went live. Subscription revenue came next, as the waitlist was released in order with a VIP tier offered as an upsell.
Treat it as a "validation-methodology textbook", not a product direction. 200 landing pages × Meta ads — your automated site-building pipeline can generate that in half a day; this validation system and your infrastructure are a match made in heaven.
Paste into Claude Code / Codex and get a working version of this product end-to-end.
Two production-grade prompts per idea — a full build spec and an SEO growth plan — plus 10 tailored build specs a week. Members only.
Join the Unicorn Club — $5/moCancel anytime, in one clickEvery claim traces back to a listed source — open any of them and check the number yourself. Timestamps jump to the second each figure is said.
Members read the exact words said here
0:54 · hear it in the videoThe exact line behind each figure is in the Unicorn Club. Read the quote, then jump to the second it's said. Join the Unicorn Club
Data credibility: Interview self-report (the founder has a $500M crowdfunding track record). All revenue figures are founder-reported or estimated by the video creator; none are audited.
The rule after launch was mechanical: every dollar that came in went straight back into ads, so scaling stopped depending on outside money as soon as return on ad spend held. Three slower loops built up around it. Reddit was worked as a content channel rather than a promotion one — long posts about task switching and what constant inbox interruption does to focus, with the product surfacing as one of the answers. Referrals turned customers into a channel: one free month for every person they bring in who signs up, and because the exchange is not a sales pitch, those users start behaving like part of the company. The feature queue became its own loop — anyone demanding a feature was offered a jump up the list in return for paying a year up front, which separates a hundred loud requests from the ones worth building and pulls cash forward at the same time. Rewriting onto their own platform and their own model then unlocked Google's CASA approval, turning Gmail access into a single approval click instead of a workaround.
This story contains no long silence, and that is the design. Every stage is built to return a number fast enough that nobody has to guess: if a page does not convert, find the single biggest problem on it and try again rather than staring at it for weeks. Where the signal came back negative, it was obeyed. Support agents were expected to be the biggest market and were not — the cycle to sign them up ran too long — so that segment was dropped, while B2C Gmail users nobody had planned for turned out to be the real demand. The WhatsApp and Instagram integrations the team was excited about went the same way once small batches of waitlist users were asked and answered with only two things: reply quality good enough that they do not have to edit it, and security. The harder stretch is the one he describes from earlier companies — hiring cheap out of old habit, then managing those hires, then becoming an employee inside his own business, and keeping a bad fit a full year. The rule he took from it: hire A players early, even before paying yourself, and fire quickly.
[Video] While the product didn't exist: (1) Used Lovable to batch-generate 200+ landing pages mapped to ~15 main ICPs plus sub-segments (solo-practice lawyers vs. law-firm lawyers vs. insurance-focused lawyers), each page with dedicated visuals and copy; the logo was two leaves drawn in PowerPoint — 5 minutes, done. (2) Meta ads with a launch budget under $5,000, manually targeted with precision by interest + age + region; creatives = solid-color background + an ultra-short CTA + emoji ("📧 too many → 😊"); the core metric was CPL, driven down to the cents level; retargeting only aimed at visitors who actually read the whole page. (3) Three-tier validation: waitlist (queue numbers starting at 1000 to manufacture buzz) → pay a $5 deposit to jump into the first 100 ("signing up is just being polite; paying is commitment") → refer 5 people, get 2 months free. (4) For segments with low payment rates (European users are wary of credit cards), an added "how I use it myself" education email sequence leveled the payment rates back up. Before validating, no Google Ads (expensive, and searchers want to use the thing immediately — a bad fit for a waitlist).
[Video] Two course corrections: (1) He assumed the biggest market was customer-support agents — wrong (sales cycles too long); instead B2C (Gmail users who hate writing email) had the demand — discovered through the multi-landing-page tests. (2) He wanted WhatsApp/IG integrations; after asking waitlist users, they only cared about "reply quality" and "security" → all cut. Lovable hit a wall in three places (write these down): security wide open, couldn't pass Google's CASA review (mandatory for Gmail authorization), and reply quality drifted with ChatGPT versions → he found an AI-developer co-founder to build their own platform + own models. The annual-plan play: "Want your feature bumped up the queue? Prepay a year" — which doubles as a second round of demand validation ("100 people shouting for it but nobody prepays a year = it's not that important").
[AI] Use this as a validation-methodology textbook, not a product direction (AI email assistants are already crowded). 200 landing pages × 15 segments takes other people two weeks; your automated site-building pipeline generates it in half a day — this system and your infrastructure are a match made in heaven. In practice: for any new idea, first spend $100-500 on Meta ads across 5-10 audience variants, validate with $5 deposits, and only write the first line of product code once it clears the bar. A reminder: his speed is built on $500M of launch-operations experience — learn the method, don't benchmark the timeline.
That was one of the 50 breakdowns open to everyone. The Unicorn Club opens the other 1,285 at the same depth — plus pain-point filters and 10 build specs a week.
Join the Unicorn Club — $5/moCancel anytime, in one click