Validation is how you find out whether anyone wants your product before you spend money making it. Here is the ladder of tests from cheapest to strongest, what a real demand signal looks like, and the fake signals that fool first-timers.
The short answer: validate by asking people to take a real action, starting with the cheapest test and climbing toward the strongest. Talk to target buyers, put up a one-page site with an email capture or a "reserve yours" button, drive a little real traffic to it from where your buyers already gather, and — the strongest test — try to take genuine pre-orders. Watch whether strangers, not friends, actually sign up or pay. The entire point is to buy a demand signal cheaply now so you do not gamble a large, non-refundable tooling spend on a product nobody wants.
Making a physical product has a hard, expensive, mostly non-refundable step: building the tooling. Once a mold is cut, that money is spent whether the product sells or not. Validation is the cheap insurance you buy before that: a few hours and a small budget to learn whether real people want this, while changes still cost nothing. Tooling first and testing demand later is a common and costly pattern. Flip that order and most of the risk disappears.
Think of your spending as reversible first, irreversible last. A landing page, a 3D-printed prototype, and a short sample run are cheap and reversible — you can change everything. Hard production tooling and a minimum order quantity (MOQ) are the irreversible end. Validation lives entirely in the reversible zone. And a trap manufacturing partners see often: a demand signal is not a manufacturability signal. A product people clearly want can still fail on design-for-manufacture, material choice, or part count — so once demand looks real, get the design reviewed before you commit to tooling, not after.
Think of validation as rungs. Each one costs a little more effort but gives a stronger signal. Climb until you are convinced — or until a rung tells you to stop.
Rung 0 — Desk research (free, about an hour). Do this first. Before you talk to anyone, read the free signals: check search demand (Google Trends and keyword interest for the problem and existing products); mine competitor reviews (the one- and two-star reviews on Amazon or retail sites show you the exact unmet complaints and roughly how many people share them); do an existing-product price check (what do current solutions cost — is there room?); and see whether similar products already sell and restock — an item that repeatedly goes out of stock and returns is selling, which is proof money changes hands. How to read the signals: a flat Trends line usually means steady, mature demand (fine — not every product needs to be trending), a sharply seasonal one means plan your launch around the peak, and a competitor with hundreds of reviews but the same one-star complaint repeated is proven demand plus a visible gap you could fill. If an hour of this comes up empty — no searches, no reviews, no one selling it — that is your cheapest possible "no".
This is where most validation goes wrong. A real signal is someone taking an action that costs them something — money, or at least their email and attention — and coming from your actual target market. A fake signal feels just as good but predicts nothing: a friend saying "great idea," social-media likes, or a stranger saying they "would totally buy it" when you never asked them to do anything. The test is simple: did the person do something, or just say something? Saying yes is free, so it means little; doing something — signing up, reserving, paying — costs them, so it counts.
It also helps to grade signals rather than treat them as pass/fail, because they form a gradient from weak to strong: an email click → an email sign-up → a sign-up with a note describing their own version of the problem → a reply to a follow-up → a small deposit → a full pre-payment → a repeat order or a referral. The further down that list a person goes, the more real the demand. A few people at the deep end outweigh a crowd at the shallow end.
Ambiguous cases come up — have a rule. Sign-ups from a paid ad still count, but weight them by how tightly the ad was targeted; broad, cheap traffic inflates weak signals. And when you get a null result (almost no one acts), do not just quit — diagnose which of three things failed: the audience (you reached the wrong people), the offer (the benefit was not clear or compelling), or the price (they want it, but not at that number). Change one variable at a time and retest. When the signal is mixed — some real interest but below your bar — treat that as "not yet," iterate on the weakest of those three, and test again rather than tooling on hope.
Decide, in advance, what result would make you proceed and what would make you stop — then hold yourself to it. If you wait until after you see the numbers, it is too easy to rationalize a weak result because you are emotionally invested. Rather than hand you a magic number (which would be made up), here is how to derive your own:
The point is not the exact figures — it is that you set the line, in writing, before you look, so the result decides for you instead of your hopes deciding.
Once you have a real signal, the next steps are designing for manufacturing and getting an honest cost, then deciding whether to license or manufacture. Validation earns you the right to spend on those steps with confidence. See how to turn an idea into a product for the full path and what it costs to manufacture to pressure-test the economics.
Once people want it, find out what it takes to make. Send Jack a sketch, a sample, or a description — he reviews it himself and gives you an honest manufacturing read. NDA first, no job too small.
Book a 30-min callWhat it costsValidate in order of increasing commitment, cheapest first: talk to target buyers, put up a one-page site with an email capture or a reserve button, drive a small amount of real traffic from where your buyers gather, and — the strongest test — try to take genuine pre-orders. Watch whether strangers, not friends, actually raise their hand or pay. The goal is to see real demand before you commit to tooling, because a signal you buy cheaply now can save a large, non-refundable tooling spend later.
The strongest signal is someone paying, or committing to pay, money — a pre-order beats an email sign-up, which beats a like or a 'that's cool' comment. Real signals come from strangers in your target market, not friends and family who want to be kind. Look for people taking an action that costs them something (money, or at least their email and attention), repeated across enough people that it is not a fluke.
Compliments ('great idea!'), encouragement from friends and family, social-media likes, and people saying they 'would definitely buy it' without being asked to do anything. These feel good but predict nothing, because saying yes costs the person nothing. Treat any signal where no one took a real action — money, a sign-up, a reservation — as entertainment, not evidence.
There is no universal magic number, because it depends on your price, market and margins. A sensible honest bar for a first-timer: if you cannot get even even a modest number of genuine strangers (say, a couple dozen as a rough example, not a magic number) to sign up or reserve after showing the product to the right audience, treat that as a warning and rework the idea. Real pre-orders, even a handful, are a far stronger yes than many free sign-ups. Decide your own threshold in advance so you are not tempted to move the goalposts.
Not always. You can validate the core demand with a clear description, renders or mock-ups before building anything. A prototype strengthens later validation — it makes the product real for buyers and for a potential licensee — but you can and should test whether anyone wants it before spending on a prototype or tooling. Validate the demand first, then invest in making it real.
Yes — a clear description, renders or mock-ups are enough to test demand. Build a prototype after you have a signal, not before.
Much better. Money, even a small refundable deposit, is the strongest proof of demand; a free email is a weaker (but still real) signal.
That is a win, not a loss — you just saved the cost of tooling a product nobody wanted. Rework the idea, the audience, or the offer, and test again.