Dreydrey9000 20-Minute Course Builder
What This Is
Ryan Magin's exact system for building a $29 digital course product in one afternoon. Not theory — this is the process that built Speed to Post, which sells 5-7 copies/day at $29 with a 20-30% conversion rate on a sales page that "looks like it was made in Microsoft Paint."
Source: Ryan Magin's workshops, team strategy meetings, and product recordings (Feb 2025).
When To Use This Skill
- the user says "build a course" / "make a product" / "I want to sell something" / "low ticket"
- A client needs to create their first digital product
- Someone has expertise but no product to sell
- Anyone asking how to monetize their audience with a simple offer
Two Tracks: The System vs. The Voice
This skill has two distinct parts. Both are valuable. One is usable immediately, the other requires input.
The System (Ryan's Verbatim Formula)
"I'm talking 29 bucks, I'm making like four or four or five hundred dollars a day right now... fucking speed the post a 20 minute video that I made it's 29 bucks chat GPT. He wrote the sales copy. It took me about three hours trying to figure it out. It's selling like five to seven a day. It's $29 and has a $19 order bump. They're all buyers."
"And I made it, the product's 18 minutes long. It's this one long loom video of me sharing my screen with some slides on it, talking about stuff."
Core Beliefs (Install These)
- A 20-minute Loom video IS a product
- ChatGPT builds the outline AND the sales page — you just talk
- One problem = one product = one afternoon
- $29 from strangers beats $0 from perfectionists
- You don't need modules, you need a message
- The sales page should look like it was made in Microsoft Paint — and that's why it converts
- "Having something to sell eliminates your need for ideas"
- "Your product doesn't need to be perfect. It needs to exist."
- The internet doesn't care about perfect products. The internet rewards people who ship.
Real-Time Research (Perplexity Integration)
This skill uses the Perplexity MCP server to research real-time topics, trends, and content opportunities. This powers the FORMAT layer — especially reaction videos and POV content.
When Research Triggers Automatically
FORMAT 2 (Reaction Videos): When generating reaction video ideas, the skill searches Perplexity for:
- Trending articles, tweets, and takes in the user's niche
- Controversial opinions or bad advice getting attention right now
- News or events their audience cares about today
Output includes: the specific article/tweet/post to screenshot, what to say about it, and the angle to take (agree loudly or disagree loudly).
FORMAT 3 (POV / Screenshot Content): When generating screenshot or "us vs them" ideas, search for:
- Popular posts with bad advice the user disagrees with
- Competitor content the user can counter
- Industry trends the audience has strong feelings about
Daily Content Research: When asked "what should I post today?" or "give me today's content":
- Search Perplexity for what's trending in their niche RIGHT NOW
- Cross-reference with their psychographic map (Layer 1)
- Match to format types (Layer 2)
- Write the scripts in their voice (Layer 3 + voice profile)
- Tell them exactly what to screenshot, what to react to, what to say
How To Trigger Research
- "What should I react to today?" → Perplexity search for trending takes in their niche
- "Find me something to screenshot" → Search for bad advice or hot takes to respond to
- "What's trending in [niche]?" → Real-time trend scan
- "Give me today's 3 posts" → Full research + scripting for all 3 daily formats
- Any content generation mode will auto-search when real-time relevance matters
How To Use This Skill
MODE 0: "Onboarding / Voice Capture" (Run First)
Trigger: "onboard [name]" / "capture voice" / "set up for [name]" / first time using skill for a new person
Walk through the intake process above. Steps:
- Ask what materials are available (recordings, reels, content, docs)
- Transcribe and process all provided recordings
- Run the 20-question interview (live or from recording)
- Pull and analyze their Instagram content
- Review their existing sales pages and copy
- Read their onboarding doc (if they have one)
- Build the voice profile:
voice-profile-[name].md - Test it: generate 3 sample CTA scripts and ask "does this sound like you?"
- Refine until they say yes
Output: A voice profile file that all other modes reference. Stored in the skill directory or user's project folder.
MODE 1: "Build My Course" (Default)
When triggered, walk through all 9 steps in order. Ask one question at a time to extract the user's one problem, expertise, and framework. Then generate:
- The one-problem statement
- The ChatGPT prompt to generate their outline
- Loom recording instructions
- The ChatGPT prompt for their sales page
- Sales page humanization checklist
- SamCart setup steps
- ManyChat configuration
- Order bump idea (next logical problem)
- First CTA video script
MODE 2: "Sales Page Generator"
When the user has a transcript, generate the sales page copy using Ryan's formula:
- Under 500 words
- Headline + subheadline + story + bullet points + price + CTA
- Written from their transcript so it sounds like them
- Include spots for iPhone Notes images
MODE 3: "Order Bump Architect"
When the user has a product live, help identify and build order bumps:
- What's the NEXT problem after they use the main product?
- Same format: slides, Loom, 15-20 min
- Price at $19 (easy add-on)
- Stack up to 5 bumps (Maria Wendt method)
MODE 4: "CTA Script Writer"
Generate daily CTA video scripts driving to the product. Requires voice profile loaded. Two formats:
Format A: Full CTA Reel (60 sec)
- Hook + value + CTA + trigger word
- Filmable in 10 minutes on iPhone
- Follow Ryan's CTA structure (see verbatim script above)
- Written in the person's actual voice (from voice profile — vocabulary, cadence, energy level)
Format B: Big Brag Pitch Reel (15-30 sec)
- Maria Wendt's formula: Big Brag + Only Next Step
- Lead with a specific result, end with one CTA
- No explanation. Just result → action.
- Must use THEIR specific results and numbers, not generic ones
MODE 6: "Pinned VSL Builder"
Build the Jason Capital-style VSL to pin to Instagram profile:
- Structured pitch video (2-5 min) following JC's framework
- Hook → pace → credibility → problem → solution → offer → close
- Ends with ManyChat trigger word
- Pin to position 1 on grid — this is the 24/7 salesperson
MODE 5: "Product Ladder Planner"
Map out the full product ladder:
- What problem does product 1 solve?
- What's the next logical problem?
- Build the sequence: each product feeds the next
- Price progression: $29 → $19 bump → $29 next product → higher tier
MODE 7: "Today's Content" (Perplexity-Powered Daily Engine)
Trigger: "what should I post today?" / "give me today's content" / "what should [name] react to?" / "find me something to screenshot"
This mode uses Perplexity to research real-time topics and generates all 3 daily posts with specific, actionable direction:
- Research phase — Search Perplexity for trending topics, articles, tweets, and takes in the user's niche. Cross-reference with their psychographic map.
- Post 1 (YAP Reel) — Pick the most relevant trending topic → write a script in their voice (from voice profile) with their angle on it
- Post 2 (Reaction Video) — Find a specific article, tweet, or bad take to react to → provide the exact link/screenshot + what to say + whether to agree or disagree and why
- Post 3 (POV/BTS/Screenshot) — Find a rallying opportunity → tell them exactly what to screenshot or film for the "us vs them" moment
Output for each post includes:
- The hook (first 3 seconds)
- What to say (in their voice, not template language)
- What to screenshot or show on screen (with link if applicable)
- CTA or no CTA (and which trigger word if yes)
- Which funnel tier it serves (top/middle/bottom)
Can also be used for:
- "Find me 5 things to react to this week" → batch research
- "What's my audience talking about right now?" → trend scan
- "Give me screenshot content" → find specific posts/comments to respond to
Full Specification
Complete details, decision trees, protocols, and implementation specs: references/full-details.md