GuideBeginner

Product Discovery and Prototyping

Teaches how to validate a bet with real users before building it. Building is the most expensive way to learn; this guide buys cheap evidence about risky assumptions before spending weeks of engineering time. It covers continuous discovery as a habit, not a one-time phase; interviews that don't lie: asking about past behavior, not future intentions, following the Mom Test approach; the opportunity solution tree for mapping the problem before jumping to a solution; turning an assumption into a falsifiable hypothesis and designing the cheapest test that can refute it (the riskiest assumption test); prototyping with AI: with a coding agent, a working prototype went from a project to a task, so the guide teaches when that prototype helps and when it misleads, how to use it as a spec, why you throw it away instead of shipping it, how to test it with users, and how to avoid false validation, prototype debt and real data in a prototype; avoiding confirmation bias and the say-do gap; and synthesizing findings (also with help from an LLM, checking every quote against the source) to decide whether to continue, pivot, or kill the bet. It picks up exactly where the Product Thinking guide left off: Mercado's `recommendations` bet, with its riskiest assumption already identified but untested. All the decision logic (which test to pick, how to weigh evidence, how to update confidence) runs as executed Node.js code. It closes with a two-gate capstone: a fake door measures demand and a working prototype built with a coding agent measures value, all the way to a discovery decision.

lessons
64
modules
8
available in
English · Spanish
certificate
Yes
access
Free
Outcomes

What you'll be able to do

  • Understand why building is the most expensive way to learn, and adopt discovery as a continuous habit, not a one-time phase
  • Interview without getting lied to: ask about past behavior, not future intentions (the Mom Test)
  • Map the problem space with the opportunity solution tree (outcome → opportunities → solutions → experiments) before jumping to a solution
  • Turn an assumption into a falsifiable hypothesis and design the riskiest assumption test: the cheapest one that can actually refute it, run with `pickTest`
  • Prototype with AI: choose between paper, clickable, fake door, wizard-of-oz and a working prototype based on the question, knowing that with an agent the fidelity versus cost curve flipped
  • Treat the prototype as a spec and as throwaway code, test it with users by measuring behavior, and avoid its risks: false validation, debt and real data
  • Avoid the biases that ruin discovery: confirmation bias, leading questions, the say-do gap; weigh evidence by strength (observed behavior beats stated opinion)
  • Synthesize interview findings by counting signals, not noise, use an LLM to code and cluster transcripts without trusting a quote you didn't check, and update a bet's confidence with `updateConfidence`, connecting back to RICE's `confidence` input
  • Decide to continue, pivot, or kill a bet with evidence, applied end to end to Mercado's recommendations bet, with a working prototype built by an agent, in the capstone project
Before you start

What you need to bring

It's for you if...

  • Engineers who already identified a bet's riskiest assumption in the Product Thinking guide and now need to validate it with real users
  • Devs who want to run cheap, fast validation before writing weeks of production code
  • Anyone who's watched a team ship a "validated" feature that was only ever validated by asking people what they'd like
  • Engineers who want a structured alternative to guessing, or to waiting for a PM or researcher to hand them the decision already made

Requirements and materials

  • Product Thinking for Engineers Guide completed (or equivalent: outcomes vs. outputs, RICE, finding a bet's riskiest assumption)
  • Comfort running simple Node.js scripts (no dependencies)
  • For the final project, access to a coding agent (recommended: the Working with Coding Agents guide first); the reference prototype ships complete in case you'd rather run it without an agent
  • No prior user research or interviewing experience required: the guide teaches it from scratch
Content

The syllabus, module by module

Open any of them to see its lessons.

Common questions

What people usually ask

Start whenever you like

Reviews

What students say

These reviews are from enrolled students who completed at least 50% of the course. We moderate reviews only on content grounds (spam, offensive language, personal data), never for being critical or negative.

No approved reviews yet.

Be the first to share your experience!