all notes
2026-07-23Abhiraj Sakargaye

Should You Validate Before You Build?.

Validate demand before you pay to build: talk to 20 people using the exact pitch you'd sell with, then stay silent and read their first reaction. Add a landing page, waitlist, or pre-sale that asks for something real — money, a date, an email tied to a problem. Stop validating once someone has said yes with a deposit, a signed pilot, or 20+ unprompted requests to use it.

Talk to 20 people in the words you'd use to sell it, then go silent — the first thing they say back is the truth. A landing page or pre-sale proves the rest. Here is what counts as a real test, and when to stop testing and start building.

Talk to 20 people, in the exact words you'd use to sell it, then go quiet. The first thing they say back is the truth — flattery only shows up after that. If nobody says "take my money" or asks when it ships, you have a demo, not a validated idea, and no studio should be taking a deposit from you yet.

Most founders skip this step because it feels like stalling. Building feels like progress; talking to strangers about an idea that doesn't exist yet feels like nothing. But a scoped MVP still costs weeks and real money, and the failure mode isn't a bad build — it's a well-built product nobody asked for. We'd rather tell you that on a free call than bill you to find out.

How many people do you actually need to talk to?

Twenty is the number that keeps showing up across founder communities, and it holds up: fewer than that and one enthusiastic friend can skew your read; more than that and you're stalling for a different reason. The method matters more than the count. On a highly upvoted Ask HN thread about customer validation, one answer put it in sales terms: write your value proposition as one sentence, say it exactly as written, then go silent — "the first thing they say is the truth," and everything after that is politeness.

Watch for the unprompted version of "take my money" — a specific date, a specific dollar figure, a request to pay now. Polite interest ("sounds cool, let me know when it's out") is not that. It's the most common false positive founders carry into a build.

Do you need a landing page, or is talking enough?

Talking gets you the problem. A landing page or waitlist gets you the first proof that strangers, not just your network, will act on it. Both Airbnb and Dropbox validated before they built the real thing: Airbnb's founders rented out air mattresses in their own apartment before writing a booking platform, and Dropbox's founder posted a three-minute video of a product that didn't exist yet — the waitlist reportedly went from roughly 5,000 to 75,000 overnight. Neither test cost engineering time.

A landing page only counts as a test if it asks for something real — an email with a real problem attached, a deposit, a "notify me at this price." Vanity signups with no ask attached measure curiosity, not demand.

TestCostTimeWhat it actually proves
20 structured conversations$01–2 weeksThe problem is real and named, not that anyone will pay
Landing page + waitlist with a real ask$50–$300 (ads or nothing)3–5 daysStrangers act on the offer, not just your network
Pre-sale or deposit$0 (you collect money)1–3 weeksSomeone will pay before the product exists — the strongest signal there is
Concierge MVP (you do the work by hand)Your time only1–2 weeksThe workflow is worth automating, before you pay to automate it
Full MVP build$10,000–$60,000+6–16 weeksWhether the validated idea holds up at product scale

The first four rows are how you earn the right to spend the fifth row's money.

When is validation enough, and when do you actually stop and build?

Stop validating once you have a named answer to "who has already said yes with something real" — a deposit, a signed pilot, a pre-order, or 20+ unprompted "when can I use this" responses from people outside your existing network. Short of that, more validation is cheap; a wrong MVP is not.

Validation isn't a permanent state either. Once you've cleared that bar, the next real cost is a scope document — a one-page brief with the problem, the user types, and a measurable success condition — because that's what turns "people want this" into a build a studio can quote accurately. We wrote up what belongs in that one page once you're past the validation gate.

Distribution is the other test people skip. A well-known reply on that same Ask HN thread said it plainly: if you don't have a way to get people to test the idea, you don't have a way to get customers once you've built it. If your validation plan is "I'll figure out marketing after launch," that's a second unvalidated idea stacked on the first one.

The one rule that decides it

Validate the demand before you spend on the build, and validate the distribution before you spend on the demand. We ask what a prospective founder has already tried — conversations, a waitlist, a pre-sale — before we scope anything. If the honest answer is "nothing yet," we say so, and the first call turns into a validation conversation instead of a quote. That's a shorter call, and a cheaper one.


Heuristics

If you've cleared the validation bar and want a real number for the build, what an MVP costs in 2026 has the tier breakdown.


Written 2026-07-23 by Abhiraj Sakargaye.

FAQ

Questions this usually surfaces.

How many people should I talk to before building an MVP?
Twenty is the number that holds up across founder communities — enough that one enthusiastic friend can't skew the read, few enough that it's not a stalling tactic. Pitch the idea in the exact words you'd use to sell it, then go silent. The first unprompted reaction is the signal; anything said after that is politeness.
Do I need a landing page or a working prototype to validate an idea?
A landing page or waitlist is enough, as long as it asks for something real — an email tied to a specific problem, a deposit, or a pre-order — not just curiosity. A working prototype tests usability, which is a different question than whether anyone wants the product at all. Save the prototype for after demand is proven.
When do you stop validating and actually build the MVP?
Stop once you have a named answer to who has already said yes with something real: a deposit, a signed pilot, a pre-order, or 20-plus unprompted 'when can I use this' responses from people outside your own network. Short of that bar, more validation is cheap and a wrong MVP build is not.