Founder Playbook · Starter Story
8 tactics from Loick
My Apps Failed Until I Discovered This Playbook ($35K/Month Apps)
Watch the full episode“We started with a newsletter that validated the initial traction then with a no code app at the time that validated the product market fit and then a real app that I've developed to really scale.”
De-Risk Validation With A Newsletter, Then A No-Code App, Then The Real App
Loick's repeatable playbook follows three escalating commitment steps: a newsletter to prove people care, a no-code app to prove they will use it, then a real engineered app to scale. He used this exact sequence on Infuspy, Mania, and Drop Magic — each hit $35K+ MRR in their first month using the same ladder.
“I use Supabase for database Next.js and Vercel for front end and hosting Sentry for debugging Trigger Dev for big compute cursor and claude code to actually develop the solution Nanobana to actually generate images with AI and then Gemini and OpenAI for the copywriting.”
Ship MVPs With Supabase Plus Next.js On Vercel And Cursor Plus Claude Code
Loick's exact MVP stack for repeatable SaaS launches: Supabase plus Next.js on Vercel, with Sentry for debugging and Trigger.dev for heavy compute. He pairs Cursor and Claude Code for development, routes AI image generation through Nanobana, and uses Gemini and OpenAI for copy. The same stack every time, so build velocity stays fixed across launches.
“with Batistan which is the co-founder of Dram Magic we knew that it would secure 10K MMR because he had like this impact and with 10K MMR you can actually now spread with other creators you can actually do some ads”
Give A Power-User YouTube Creator Equity To Lock In Distribution From Day One
Loick brings a YouTube creator on as equity co-founder to collapse the two biggest early risks — product and distribution — into one person. The creator guarantees baseline MRR (Drop Magic locked in ~$10K from Batistan alone), surfaces product feedback as a power user, and ships free iteration on scripts, funnels, and pricing until the unit economics are proven enough to replicate with paid creators.
“to find a really good creator I actually have this four-step uh system... engagement rate uh so basically it's like views uh divided by followers it should be over 10%... a minimum of like 100 comments per video... did the creator worked at least with uh the brand three times... I have a seven email campaigns... always lowable the initial offer by 30% and then offer two package”
Find Creators Through A Four-Step System: Search, Vet 10% Engagement, Multi-Email Blast, Lowball Two Packages
Loick's repeatable creator system: (1) search YouTube keywords and study competitor partnerships, (2) vet on engagement above 10% and 100+ comments plus 3+ repeat brand deals as a profitability signal, (3) hit them with a 7-email sequence plus Instagram and Twitter DMs, (4) lowball by 30% and offer two packages — higher upfront with lower commission, or lower upfront with higher commission. Always start with a one-video test, break-even check, then bundle 3-4 videos.
“TikTok is really good form for like brand awareness but it doesn't convert whereas YouTube is just machine for conversion like for ourself I think more than 60% of our new customers come from YouTube”
Bet On Long-Form YouTube Because It Converts, Compounds, And Repurposes Down Into Shorts
YouTube is Loick's only channel bet for four compounding reasons: predictable minimum views (no virality lottery), long-form runtime that lets the creator actually demo the product, evergreen conversion (3-year-old videos still close customers), and the option to chop long-form into shorts and ads later. TikTok shorts go the other direction — awareness without conversion — so he refuses to start there.
“we pay like 1.5K because we actually offer uh five AI images for every free user so that scales pretty much”
Burn $1,500 Of LLM Spend On Free-Tier AI Images So Users Feel The Core Value Before Paying
Drop Magic gives every free user 5 AI-generated images as part of onboarding, which costs roughly $1,500/month in LLM bills. It's the biggest variable cost but lets users experience the core value before paying — and that activation cost still pencils out at 70-80% margins because conversion lifts on the back of the free demo.
“you have server and compute you have like 100 bucks per month and then for all the AI so we use a lot of LLM of course so we pay like 1.5K because we actually offer uh five AI images for every free user so that scales pretty much... I think the sweet spot will be between 70% and 80% margin”
Hit 70-80 Percent SaaS Margins On $100 Infra Plus $1,500 Of AI Passthrough
Loick's cost structure is brutally lean: ~$100/month infra, ~$1,500/month AI (driven by giving free users five AI images as a top-of-funnel hook), and the rest is ad/creator spend. Mania ran at 80% margin; Drop Magic sits a touch lower only because they're aggressively testing. Bootstrapped SaaS economics still survive heavy LLM usage if you cap free-tier generations.
“just don't get emotionally attached to a project just use simple deadlines and milestone as the only source of truth to carry on or not i've wasted so many years to just you know believing in something that had no traction just because I was emotionally attached to the logo or the project or the team members”
Let Deadlines And Milestones Be The Only Signal Because Ego Kills More Startups Than Bad Code
Loick names emotional attachment as the single biggest tax on his career — he kept resuscitating products because he loved the logo, team, or idea, not because the numbers said so. His prescription: pre-commit to deadlines and milestones, and let traction (or its absence) be the only input to the kill-or-continue decision.