Idea Validation

Source
TL;DR	Don’t validate with free email signups — validate with money or a dated commitment. Make customers prove intent with a wallet or a calendar before you write code, and stay willing to let the idea die.

What it means

The #1 cause of startup death is building something nobody wants. Code is cheap now (vibe-coding); the bottleneck is aiming.
The trap is soft validation — free waitlists, “yeah I’d use that,” viral demos with no purchase. Feels like demand, measures nothing.
Real validation forces a costly signal: a credit card, a pre-order, a committed date — before the backend exists.

Money beats email.

Ask people around you about their problems, then “would you pay $X to solve this?” — and have them pay before you build (Nadav got $20K pre-build); the first payer is your perfect ICP template (i-asked-20-founders-how-to-find-a-1m-startup-idea).
Aayush: ship a paid buy button day one — “real learning only happens when money is exchanged” (how-i-grew-my-saas-to-150k-year-with-reddit-and-seo).
Premium Tier on Day 1: Josh (app acquirer) recommends launching with a premium tier from the first day to validate early whether users see actual value in paying; if they don’t, either adjust price, expand the premium scope, or pivot immediately (build-7-figure-app-josh).

Validate marketability, not just demand (for consumer apps).

Cravotta validated Puff Count before building by proving the problem went viral — listing quit-vaping TikToks and their view counts in a sheet. A great but unmarketable app is worth $0; check Sensor Tower revenue + Google Trends + social virality (how-i-build-apps-that-print-30k, app-market-research).
Angus: run paid Google Ads at launch purely to test demand; kill them if unprofitable (i-make-40k-month-with-this-one-website).
Your Average Tech Bro’s first gate is even stricter: “can I market it?” If the answer is no, don’t build — AI makes technical feasibility cheap, but marketability still decides whether the app matters (my-step-by-step-framework-for-validating-app-ideas-2026).
Adam Lyttle adds two harder consumer-app filters: inspect competitor revenue/download estimates before building, and ask whether the target user can realistically afford a subscription. Demand without purchasing power is not a business (adam-lyttle-will-your-app-idea-even-make-money-spy-on-competitors-with-this-shor).
John Rush: 100 signups → pre-sell a 90% discount → 5 sales = build it, else don’t (i-make-3m-year-from-20-different-products).

A dated commitment beats a card.

Brian Shin wrote no code until he had “10 events with an actual date and an actual commitment” (i-built-a-20k-month-app-in-83-days). Demand you can’t fake.

Waitlists are useless — get a payment or a proven format instead.

Dan (Massive): ~0.5% of a 40–50k waitlist converted; “use that time to get people to pay, even for a small version.” Better: test the message with influencers who already get consistent views, or run cheap ads on a prototype (massive-tinder-for-jobs-2m, roger-chen-number-1-app-twice).
Native store pre-orders are a better mobile waitlist, but still not proof of paid demand: Apple/Google can concentrate launch-day installs, not prove retention, habit, or subscription conversion (app-store-preorders, apple-app-store-preorders, google-play-preregistration).
Adam’s App Store angle adds one more caveat: some launch-week monetization strategies exploit Apple’s temporary visibility boost, but that is operationally narrow and not proof of a durable business by itself (adam-lyttle-this-faceless-app-developer-makes-20k-month-from-the-app-store-boost).
The strongest organic signal is unprompted pull — friends asking “can I use this? can I give it to people?” (your-move-ai-30k-side-business).

Talk to humans before you build a pixel.

Ombberto: interview 5–10 target users about the problem, never reveal you’re building a solution — “read the Mom Test” (my-app-made-120k-in-24-hours).
Anya: mined negative reviews of mental-health apps to find the gap (in-the-moment relief), built around one SOS button (this-app-made-over-1m).
Perfect Interview’s B2B pivot came from one career coach asking to white-label the tool, then a few more coaches saying yes — “build for persons, not personas” (why-im-pivoting-my-saas-from-b2c-to-b2b).
Chris Raroque adds a goal-based filter: if the goal is learning or a small side income, personal workflow pain plus tiny scope can be enough; if the goal is $5-10K/month or a real company, validate market size, reachability, and willingness to pay much harder (i-built-3-apps-heres-how-i-pick-winning-ideas).

Cloning is validation by proxy.

If a competitor posts MRR screenshots, the market is proven — copy the demand, not the risk (Mike, i-built-3-saas-apps-to-200k-mrr). See the-clone-strategy.
Corollary from 20 founders: an idea with no existing solution is a red flag, not whitespace — validating raw demand is too hard (i-asked-20-founders-how-to-find-a-1m-startup-idea).
Or skip product risk entirely: run an agency/service first, spot the recurring problem, build one feature, sell it back to clients you already have (distribution-cheatcode-1-podcast-150-posts).

The hardest part is emotional, not mechanical.

Jordan: validation is “willingness to let your idea die,” not a framework (i-made-1-5m-from-an-app-you-ve-never-heard-of).
Founders who pivoted fast on bad signal won: Ethan scrapped a coach-facing MVP for athlete-direct (i-left-college-to-build-this-niche-app-now-i-make-20k-m); Prop GPT rebuilt over 4 months after a 13% trial-to-paid rate (how-we-built-a-30k-month-mobile-app).
Validation you ignore is worse than none.

Do this, not that:

Ship a paid buy button or launch with a premium tier on Day 1 — not a silent free waitlist.
Use native store pre-orders to concentrate already-proven mobile demand — not as validation by themselves.
Collect dated commitments — not vague interest.
Interview about the problem — never pitch the solution.
Match validation depth to ambition — don’t use “I have this pain” as proof of a venture-scale market.
When the numbers say no, kill it — don’t protect a beloved idea for a year.