smoke-test-launcherlisted

Design and launch demand smoke tests — fake-door pages, preorders, waitlists, and paid-ad probes — with the pass/fail threshold and minimum sample size signed before launch, then debrief results honestly. Use before building anything to test demand, when a landing page or waitlist experiment is planned, or when the user says smoke test, fake door, test demand before building, should we build it, or validate interest. Not for testing product value or usability (a smoke test measures demand, not the product) or for running full A/B experiments.
karozi/awesome-product-management-skills · ★ 2 · Testing & QA · score 56

Install: claude install-skill karozi/awesome-product-management-skills

# Smoke Test Launcher A smoke test asks one question: does anyone want this enough to act? It measures demand for an offer that does not exist yet — cheaply, honestly, and with the verdict written before launch so nobody can move the goalposts after. ## The Rule The threshold and the sample size are signed before the page goes live. A test "designed" after results arrive is a press release for a decision already made. ## Modes ### Draft mode (default — user has an idea to test) 1. Read `references/smoke-test-types.md` before choosing a format. 2. Extract the hypothesis: who wants what, badly enough to do what. If the user cannot state the action that counts as demand (signup, preorder, waitlist join, booking), ask once — the action IS the test. 3. Choose the format: fake-door page, preorder, waitlist, or paid-ad probe — matched to whether the risk is interest, willingness to pay, or message. 4. Write the page: headline (the promise in their words), one paragraph of mechanism (why it can deliver), the CTA (the demand action), and nothing else. Every extra element dilutes the signal. 5. Copy `templates/smoke-test-plan.md` and fill every bracket, including the source of traffic and its expected quality. ### Threshold mode (user has a page but no signed verdict criteria) 1. Set the pass threshold from the traffic source's baseline, not from optimism: [X]% conversion from [N] qualified visitors, with N sufficient for the number to mean anything. 2.