01 The problem
Aspiring PMs — people like me, mid-career, switching in — face a wall of content: articles, frameworks, YouTube courses. You can consume all of it and still freeze the first time someone asks you to write a PRD or prioritise a backlog.
The gap is between knowing about product management and being able to do it under pressure. Passive content doesn't close that gap; practice with feedback does. This is the single clearest thing a decade of teaching taught me — students who only watch lectures don't improve; students who attempt problems and get corrected do.
You don't learn Chemistry by reading the textbook cover to cover — you learn it by attempting problems, getting them wrong, and understanding why. PM is no different. PM Quest is that principle turned into a product: attempt, get graded, learn from the gap.
02 The design bet
The bet is that PM learning should feel like Duolingo, not a textbook — short, active, repeatable units with immediate feedback, rather than long passive lessons. The hard part isn't the gamification; it's the feedback. Multiple-choice is easy to grade but tests recognition, not thinking. Real PM questions are open-ended — "how would you prioritise these three features?" — and those need judgement to assess.
03 How it works
Bite-sized units
Concepts delivered in short, active lessons rather than long-form reading.
Case mode
Open-ended PM scenarios — prioritisation, metrics, tradeoffs — that mirror real interview and on-the-job prompts.
AI feedback
Groq-powered grading evaluates the reasoning in a free-text answer, not just a right/wrong key.
Track & return
Progress persists per user, so learning compounds across sessions rather than resetting.
04 The hard part: grading open answers
The core product challenge is the same one at the heart of ChemIQ — how do you evaluate a free-text answer fairly and consistently? A grader that's too harsh discourages; too lenient and the feedback is worthless. Getting the grading prompt to assess reasoning quality — did they consider the user, the tradeoff, the metric — rather than keyword-matching is the difference between a real learning tool and a quiz.
- Grade the thinking, not the wording. A good answer phrased plainly should beat a buzzword-stuffed weak one.
- Feedback has to be specific. "Good job" teaches nothing; "you didn't state a success metric" does.
- Consistency across attempts matters — the same answer should get the same grade, or users stop trusting it.
05 What I'd watch next
Honestly, PM Quest is live but hasn't been through structured user testing. The questions I'd want answered:
- Does the AI grading feel fair? If learners dispute their scores, the grading rubric needs tightening — the same evaluation-first discipline I applied to ChemIQ.
- Does the Duolingo format actually drive completion? Gamification can feel gimmicky if the underlying content isn't strong. The test is whether people return.
- Is it teaching PM, or teaching how to answer PM questions? The real goal is transferable skill, not interview performance — and that's the hardest thing to measure.
06 What building it taught me
- My teaching background is the product edge. Most PM-learning tools are built by PMs. This one is built by someone who spent ten years figuring out how people actually acquire a hard skill — and that shows up in the active-practice design.
- Feedback is the product, not content. Anyone can list PM frameworks. The value is in grading an attempt well enough that the learner improves.
- The same hard problem recurs. Fair free-text evaluation showed up in ChemIQ and again here — it's becoming the through-line of what I build.