A growing collection of structured product analyses — written from the perspective of a Senior PM who wants to understand not just what a product does, but why it was built that way, and what it means for anyone competing in the same space.
These aren't reviews. They're strategic dissections.
The best PMs I've worked with share one habit: they study products outside their own domain obsessively. Not to copy features, but to internalize the decision-making behind them — the tradeoffs, the business model constraints, the moments where the team clearly chose retention over ethics, or simplicity over power.
Writing teardowns is how I stay sharp. It forces precision. You can't hide behind vague product intuition when you have to articulate why Duolingo shows you a passive-aggressive owl instead of a modal.
I also pay particular attention to how AI is reshaping each product's strategy — where it's a genuine differentiator, where it's a bolt-on, and what the AI-native competitive threat actually looks like.
| Product | Category | AI Angle | Read |
|---|---|---|---|
| Duolingo | Consumer / EdTech | AI roleplay, real-time translation threat, LLM disruption risk | → Read |
| Splitwise | Consumer / Fintech | AI-assisted expense parsing opportunity | → Read |
| Perplexity | AI-Native / Search | Answer engine vs. Google Search, publisher conflict, enterprise pivot | → Read |
| LinkedIn Job Matching | Professional Network / HR Tech | LLM-powered matching vs. Indeed, recruiter revenue threat, AI bias risk | → Read |
| Character.AI | AI-Native / Consumer Social | AI companionship vs. Replika, parasocial attachment mechanics, minor safety risk | → Read |
New teardowns added periodically. Watch the repo to get notified.
Every analysis follows the same framework — structured enough to be comparable across products, opinionated enough to be useful:
1. Product Overview & Positioning What is this product, who is it for, and what's the competitive context? No fluff — just the facts that matter for strategic orientation.
2. Core Jobs-to-be-Done The functional and emotional jobs users are hiring this product to do. This is where most product analyses go shallow. I try to go one level deeper: what's the job behind the job?
3. Key UX Decisions & Why Specific design choices, with hypotheses about the reasoning behind them. Onboarding, core loops, friction points, and the moments where you can see the PM's fingerprints.
4. Business Model Mechanics How the product makes money, how it acquires users, and what the retention engine actually is. Because features don't exist in isolation — they exist to serve a business model.
5. Recommendations & Strategic Risks If I were the PM: what would I ship next, what would I be scared of, and where is the AI-native competitor going to attack first?
I'm particularly interested in products at an inflection point — where AI has either already changed the product fundamentally, or where AI represents an existential threat to the current model.
Duolingo is a good example: its core value proposition (casual language learning) is under direct threat from real-time AI translation embedded in earbuds. How does a product survive when its functional utility becomes ambient infrastructure? That's the kind of question I want to sit with.
Going forward, I'll be prioritizing teardowns of AI-native products and products facing AI-driven disruption.
If you're a PM who does this kind of thinking and wants to exchange notes, I'm always open to a conversation.