Most no-code validation advice optimises for a feeling. Put up a page, get some signups, feel better, start building. The signups are real and the feeling is not evidence.
What follows is a sequence that produces things you can be wrong about — which is the only kind of finding worth having before you spend three months.
Decide what would change your mind
Write this down before anything else, because it is the step that makes the rest mean something.
Finish both sentences:
If ___, I will build this. If ___, I will not.
If you cannot fill in the second one, you are not validating, you are gathering encouragement. That is a legitimate thing to want and it is not what this is for.
The bar does not have to be sophisticated. "Forty people who describe the problem the way I do" is a fine bar. "Some signups" is not, because no result fails it.
Week one: find out what people currently do
Before a page, before a form, before anything: twelve conversations.
Not "would you use this". That question gets a polite yes and tells you nothing. Ask about the past instead, because the past actually happened:
- What did you do the last time this came up?
- What did that cost you — money, time, or annoyance?
- What did you try before that, and why did you stop?
- What would have to be true for you to change?
Twelve is enough to hear the same sentence three times, and hearing the same sentence three times is the finding. Write down the words people use. Those words are your landing page; the words you would have written are not.
You do not need software for this and you should not skip ahead to the part where you do.
Week two: a page, and one real question
Now a page. Not a product, not a demo — a page that describes the thing in the words you just heard, with a way to say "yes, me".
Four things it must have.
Double opt-in. Verify the address. Without it you do not know your denominator, and every number you compute afterwards rests on an unknown fraction of disposable addresses, form-filling bots, and — the moment you add referrals — people gaming a queue position. This is free on every Osokoro plan, including the free one, for exactly that reason.
One free-text question. "What do you do about this today?" is the best one. Read every answer yourself, all of them, before summarising anything. This is where the finding is, and averaging it away is how founders end up with a dashboard and no insight.
A price question, read the way the next section describes.
Source tracking. A number without a source is a number about your friends.
What a price question can and cannot tell you
A stated price is a statement, not a payment. Nobody on a waitlist has transacted, and the gap between what people say and what they do runs in one direction. That does not make the answer worthless — it makes it comparative rather than absolute.
Good for:
- Ranking. If $19 answers outnumber $9 answers three to one, that ordering is informative.
- Finding the cliff. Where willingness falls off a shelf is usually visible, and usually not where you guessed.
- Segmenting. The people who said $49 are a different population from the ones who said $9, and the difference between what those two groups wrote in the free-text box is often the most valuable thing you collect.
Not good for: a number in a spreadsheet. Osokoro shows a recommendation next to these answers, and it
carries a disclaimer saying it is not a guarantee of demand, revenue or future purchases. The
thresholds behind it live in VERDICT_THRESHOLDS in packages/domain/src/validation.ts, and the
product links to them from the page — a heuristic that hides its cutoffs is asking for more trust
than a heuristic deserves.
Week three: the test everyone skips
Write to the list. Something real: an update, a question, a rough sketch of what you are thinking.
Then count replies, not opens.
If four hundred verified people receive a genuine email and almost nobody answers, you have learned something sharper than any positive number, weeks before you would otherwise have learned it. Silence is the cheapest strong signal available and almost nobody collects it, because collecting it means sending something and risking the answer.
One detail you should have now rather than later: sending on the shared domain works on every plan. Everything includes one custom sending domain and Studio allows several; your mail keeps using the shared domain until the DKIM records you publish have been verified.
The ladder of evidence
Ranked, weakest first. Most founders optimise rung one and read the result as though it were rung six.
- A signup — somebody typed an address.
- A verified signup — the address exists and its owner clicked.
- An answer to a specific question about their current situation.
- A stated price, read comparatively.
- A referral — they put their reputation in front of somebody they know.
- An unprompted reply, weeks later.
- Money — not available from a waitlist, by construction.
The useful work is moving people down that ladder, not accumulating more of rung one.
What none of this tells you
Whether you can build it. Whether you can reach these people at a cost that works. Whether they will still want it in six months. Whether somebody else ships first.
Validation before building bounds exactly one risk — that nobody wants this — and leaves every other risk where it was. Anyone selling it as more than that is selling you the feeling.