Founder guide
How to validate a physical product before manufacturing
A staged validation process that tests the customer problem, the offer, the price, and the production assumptions before a large commitment.
Direct answer
Validate the riskiest assumption first. Confirm that a specific customer recognizes the problem, then test the offer and the price, then collect a commitment that costs the customer something, and only then increase manufacturing spend. Pre-orders against a real storefront are the cheapest strong signal available.
Write a falsifiable customer claim
Name the customer, the situation, the problem, and the proposed advantage. "People want better wellness products" is not testable, so it can never be wrong, which is exactly why it feels comfortable.
"Night-shift nurses need a shelf-stable snack that meets a specific dietary constraint" gives you an audience to interview and an offer to put in front of them. A claim you could lose is the only kind worth testing.
Interview for behavior, not compliments
Ask what the person did the last time the problem occurred. What did they buy, what failed, what did it cost them, what did they do instead? Past behavior is evidence. Enthusiasm about an imagined product is not.
Avoid asking whether they like your idea. Compliments are free to give and people give them to be kind. The useful signals are specifics: a receipt, a workaround, a complaint they have repeated to somebody else.
Test the offer before the production run
Show a clear product concept, a price, expected timing, and terms. A waitlist tests whether the message lands. A pre-order tests a much stronger commitment, and it requires accurate disclosures, real payment handling, and a credible refund process.
The gap between those two signals is where most bad launches are decided. Interest is cheap and abundant. Payment is scarce and informative.
- Track qualified visits separately from raw traffic
- Measure email signups, checkout starts, and paid commitments separately
- Record objections in the customer’s own words, not summarized
- Do not treat friends, employees, or paid testers as an unbiased sample
Run the test without building anything first
You do not need finished inventory, or even a finished product, to run a real offer test. In Dough the sequence is: describe the idea, refine one of the generated drafts until it is the product you mean, then create it and let the storefront and branding build in the background.
From there, set the price with the breakdown visible so you know the margin you are testing, set a launch goal with a target quantity and a deadline, and share the buy link. Choose pre-order mode to test payment, or waitlist mode when you only want to test the message.
What you get back is a demand signal attached to your product at your price, which is the only signal that transfers to the actual launch.
Validate production assumptions too
Demand does not prove manufacturability, and this is where validated ideas most often die. Confirm the specification, the target cost, the sample plan, minimum order quantity, lead time, quality checks, labeling, compliance, packaging, freight, duties, and the fulfillment path.
Run this track in parallel with the demand test rather than after it. A product that sells beautifully at a price the factory cannot support is a more painful discovery than one nobody wanted.
Set a decision rule in advance
Decide what evidence will make you proceed, revise, or stop, and write it down before the results arrive. Thresholds chosen afterwards are chosen to justify what you already wanted to do.
The threshold should come from your contribution margin and your production commitment, not from a conversion rate quoted for a different category with a different price and a different buyer.
A staged plan you can run this month
Week one: write the falsifiable claim and talk to ten people who fit it, asking only about past behavior. Week two: build the product concept and storefront, set a price you can defend from the cost breakdown, and put the offer in front of that same audience.
Week three: run the offer to a small, qualified audience with a launch goal attached. Week four: compare the result to the rule you wrote in advance, then either write the production specification or change the claim and start again. Both outcomes are cheap at this stage. Neither is cheap after a production run.