Launch a New Product
A one-week course that turns a launch from a leap of faith into a sequence of cheap tests: find the assumption that kills you, test it first, and only then build.
A launch isn't won by shipping — it's won by being right about the segment and the value before the budget runs out. This course makes being wrong cheap and early.
What's inside
1. RAT: kill the risk before it kills you
50 min · 2 interactive exercisesThe riskiest-assumption procedure: enumerate what must be true, rank by damage × uncertainty, test the top one for days-not-months. Includes the live-filled risk table from the workshop.
Result: Your launch's ranked risk table, riskiest assumption on top
2. An MVP is not a product
35 min · 1 interactive exerciseAn MVP is a probe — it exists to answer one question, not to be a small version 1. What your MVP actually tests, and the Lean-Startup failure mode where 'validated learning' validates nothing.
Result: The one question your MVP actually answers, stated in writing
3. Research or MVP: don't burn the budget
35 min · 1 interactive exerciseSometimes asking is 10× cheaper than building. The decision procedure: when expert interviews beat a prototype, and how to use them as a flashlight before committing engineering.
Result: The cheapest-test decision for your killer assumption
4. Design a segment-search study
50 min · 2 interactive exercisesWho to recruit, how many interviews, and the iteration loop of 6–8 that converges on a segment instead of circling. Includes the de-narrowing procedure for when your assumed audience is wrong.
Result: A recruiting plan: who, how many interviews, in what iteration loops
5. The traction map
45 min · 2 interactive exercisesLaunch discipline in one table: stages, metrics, thresholds. A red cell means stop and diagnose — not push harder. The template plus a live walkthrough of filling it.
Result: Your traction map with thresholds and the red-cell stop rule
6. Backward planning
35 min · 1 interactive exercisePlan from the launch result backwards to today. It surfaces the dependencies a forward plan hides — usually the exact ones that slip.
Result: A backward plan from the launch result to next Monday
Who it’s for
- You (or your team) can now build the product in weeks — which removed the last natural pause before spending a quarter on the wrong thing.
- Your last MVP took two months and 'validated' nothing anyone can name.
- You're staring at a launch plan that's really a build plan — no test in it can fail in a way that changes the plan.
- You need to convince a board or a boss that the launch is de-risked, with something more solid than confidence.
What you leave with
By the end of the week you hold a ranked table of the assumptions your launch stands on, know which single one kills the bet if wrong, have the cheapest test for it designed — and a traction map where a red cell means stop, so discipline survives contact with enthusiasm.
- A ranked risk table for your launch (riskiest assumption on top)
- The cheapest-test design for your killer assumption
- Your traction map: stages, thresholds, and the red-cell stop rule
- A backward plan: from the launch result to next Monday
Artifacts run on your product and are submitted in the LMS — an artifact about your business, not a certificate of completion.
What your agents get
You learn the method; your agents in Claude Code and Codex get the same algorithms as installable skills. The team that learns is you plus your agents.
Derives candidate segments and their riskiest assumptions for a new product before you've talked to anyone — the starting map the study design then tests.
Turns the validated direction into a build spec with the risk table attached — so what you hand engineering carries the launch logic with it.
Format
4–6 modules, 3–6 hours of content — done in one working week alongside your job. Exercises run on your own product; artifacts are submitted in the LMS as you go.
What it’s built on
The method behind this course is public. This course packages these canon pages into a playbook with exercises:
What you’re probably worried about
“Testing first will slow the launch down.”
The RAT procedure is measured in days: the point is finding the test that takes a week instead of the build that takes a quarter. The course exists because building has become the slow way to find out you're wrong.
“My investors/boss expect shipping, not research.”
The risk table and traction map are shipping artifacts — they're what de-risked execution looks like on paper. Teams use the exercise output directly in board updates.
“What if the killer assumption test fails?”
Then it failed in week one for the cost of a test, not in month four for the cost of the build. The traction map's stop rule is the feature, not the failure.
Take the task off your plate this week
One working week, on your own product, artifact at the end. $399 — one prevented wrong sprint pays for it.