Where Colin started
Colin was a Senior PM at Datavant. No following, no newsletter, no plan to build a brand. He was the most technical PM on his team, constantly helping colleagues understand how the platform worked, watching the same knowledge gaps slow decisions down every day. Then a mentee at a bank had the exact same problem, and that was the signal he needed. To nail his course topic, he just looked at what he was already fixing for free.
His first two courses
His first two courses
Course 1 · 2022 — Technical Foundations for Product Managers
Topic source: daily work frustrations.
His PMs at Datavant couldn't explain engineering timelines, couldn't challenge estimates, couldn't communicate how the platform worked to customers. He saw the same gap with a mentee at a bank. Seeing the same problem at two companies let him know there was a market.
Course 2 · 2024 — AI Prototyping for PMs (Vibe Coding)
Topic source: tools he was already using.
He was using Bolt and Lovable himself just two months after the tools launched. He saw a specific enterprise use case nobody was teaching: not "build a side app," but "test ideas at work, share prototypes with your team, move faster." And, he built the course before most people had heard of the tools.
"The two courses I've had that are most successful were one of each: I saw a problem with my coworkers. And I made a bet that this space would become important." — Colin Matthews, Maven Instructor Session, Feb 2026
How the first course actually got off the ground
How the first course actually got off the ground
Cohort 1: 5 students, 2 real sales
Colin kept cohort one to a true beta experiment, and didn’t see a ton of success. But he kept going anyway because he thought he could make the product better.
Cohorts 2–5: 15 → 30 students per cohort
He did zero marketing and no social posting. Instead he had an open cohort monthly, and the Maven marketplace drove organic discovery. This isn't always typical, but it worked because the topic had real search demand. The most successful courses on Maven see strong course topic fit!
Lenny cold-DM'd him on LinkedIn
He still doesn't know exactly how Lenny Rachitsky found him. His theory is that it was because of his Maven visibility. They co-wrote a newsletter piece on technical foundations for PMs, and Colin's subscriber count jumped. The credibility of being a Maven instructor made the introduction possible.
Inbound from newsletters, then B2B workshops
After Lenny, other newsletters found him (including a 150K-subscriber system design newsletter). Corporate training requests then followed, with out him ever running an ad.
The principles that made it work
The principles that made it work
Colin looked at what he was already solving for free
Both courses came from problems he was fixing at work or in mentoring. He focused on packaging what he already knew.
He counted "minimum meaningful reps," not single tries
He gave himself 4–5 cohorts before deciding if something worked. One cohort wasn't a test for him. The data only started to mean something after a few runs.
Having a job kept his content sharp
He stayed employed by choice, knowing that real work keeps course content current. The AI Prototyping course existed because he was using the tools himself.
He gave away everything and focused on selling the practice, not the information
His posts for Lenny and other newsletters are basically the course in written form. He doesn't gate knowledge. What students are actually paying for is live sessions, direct access to him for follow-up questions, and doing the work alongside others.
He did one thing first, then expanded
He ran one course monthly for over a year before adding a second. He's built SaaS tools, a Slack community, a Substack, and consulting work, but only after the course was stable. He estimates a 20% hit rate across everything he's tried, and feels that his focus raises those odds.
Two ways to find a course topic
Two ways to find a course topic
Approach A: Solve a problem you already see (lower risk, harder to size)
Your coworkers have this problem. Your friends ask about it. You've fixed it multiple times. The demand is already in front of you, and you're just not charging for the solution yet.
Approach B: Bet on a space before it's obvious (higher variance, higher ceiling) You're using a tool or working in a domain that feels early. You believe it's going to matter. Get there before the crowd and your lack of competition will be your advantage.
"There isn't really a big difference between someone you think is so smart and doing such great work, and yourself — other than maybe 6 to 12 months of applied effort." — Colin Matthews, Maven Instructor Session, Feb 2026