How to validate an ecommerce or physical product idea
The short answer
2
orders, the point where paid acquisition works or does not
Demand is rarely what kills a physical product. Arithmetic is. Work out the landed cost of one unit, then subtract shipping, returns and what it costs to acquire the buyer. If what remains does not cover a second order's acquisition, you have a hobby with inventory rather than a business.
This assumes you are selling online and buying traffic. Selling through wholesale, a market stall or an existing audience changes the acquisition term substantially and can rescue margins that fail here.
01
Five things to check before you order stock
Every check below can be done on paper, before any money is committed to inventory. That is the whole advantage available to you here, and it is routinely skipped in favour of asking people whether they like the product.
- 01Do this first
The landed cost of one unit, not the factory quote
The unit price a supplier quotes is the beginning. Add freight, duty, packaging, and the fraction of your minimum order that never sells. Founders routinely build a business case on the quoted figure and discover the real one after committing to stock, at which point the price they planned to charge no longer works.
- 02
What shipping actually costs, both ways
Delivery is expected to be free or near it, which means it comes out of your margin rather than the customer's pocket. Cost it honestly for your heaviest, most awkward variant, and include the return leg. A bulky low-priced item can be uneconomic to ship before anything else is considered.
- 03
Your return rate, by category
Returns vary enormously by what you sell. Anything sized, fitted, or judged on colour comes back at rates that can consume the entire margin, and a returned item is often not resellable. Find out what is normal for your specific category and treat that as a cost line rather than an edge case.
- 04
What it costs to acquire one buyer
If you are selling online to strangers, you are buying attention, and that price is set by whoever else is bidding for the same customer. Look at what the established sellers in your category are doing and assume you will pay more than they do at first, because you have no data and no reviews.
- 05
How often the same person buys again
This is the term that decides whether paid acquisition ever works. If someone buys once every five years, every sale must be profitable on its own, which is a brutal constraint. If they buy monthly, you can afford to lose money on the first order. Establish which world you are in before you model anything.
A product people love that loses four pounds a unit is not an early stage business. It is a subsidy you are paying strangers to accept.
02
What counts as evidence here
Interviews still matter, but they answer a narrower question than in software. Ask about purchases that already happened, because a price someone endorses in conversation is not a price they have ever paid.
Counts
- What they paid for the nearest comparable thing, with a figure
- A pre-order or deposit at your real intended price
- Someone describing what is wrong with what they currently buy
- How often they replace or repurchase this kind of item
- A supplier quote you have actually received, with freight included
Does not count
- I would definitely buy that
- A price someone names without being asked to pay it
- Interest at a price below your landed cost
- Compliments on the design from people who are not the buyer
- Social engagement on a product photo
03
Questions founders ask next
How do I validate a physical product without manufacturing it?
Sell it before it exists, at the real price, with an honest delivery date. A pre-order page collecting actual payment is the strongest signal available and costs far less than a production run. If that feels dishonest, it is worth noticing that taking money for something you have not made is exactly what you are proposing to do at scale.
How many people should I talk to for a product idea?
Fewer conversations than software needs, and more arithmetic. Ten conversations will tell you whether the problem and the price are plausible. After that the binding questions are landed cost, return rate and acquisition cost, none of which improve with a larger interview sample because none of them are opinions.
What margin do I need for an ecommerce business to work?
Enough that contribution margin after shipping and returns covers acquisition cost with room left over, which for a one-off purchase usually means the retail price is a large multiple of landed cost. Work backwards from your own numbers rather than adopting a rule of thumb, since the required multiple depends entirely on how often people repurchase.
Is a crowdfunding campaign good validation?
It is strong evidence on demand and price, since backers pay real money before the product exists. It is weak evidence on unit economics, because campaign backers are unusually motivated and you did not pay to acquire most of them. Treat a funded campaign as proof people want it, not as proof you can sell it profitably to strangers afterwards.
Should I sell on a marketplace or my own store?
A large marketplace brings buyers you would otherwise pay to reach, and takes a commission plus your relationship with the customer. Your own store keeps both and requires you to generate every visit yourself. The right answer depends on the acquisition cost you calculated, and the commission is usually cheaper than advertising at the start.
My product idea already exists on Amazon. Should I stop?
Not necessarily, but your differentiation now has to be visible in a photograph and a title, because that is all a buyer compares. Read the one and two star reviews of the closest sellers, since that is where people describe precisely what failed. If the complaints are about things you cannot fix at your cost base, that is a finding worth taking seriously.
What you get back
A look inside a real brief
Not a screenshot of an interface. This is the artifact a round produces, with every score attached to the evidence behind it. It is an extract; the full one goes further on all of it.
The idea
Booking and no-show fees for independent dog groomers
Sources read
37
Confidence
Medium
The signal is there. Go build it.
68% of 19 people confirmed they have this problem.
How we got to that number
Thirteen of nineteen described losing money to no-shows without being prompted, and six already pay for something to reduce it. The score is held below seventy because the free workaround is tolerated rather than hated, which is the switching cost you have not tested yet.
Section 1 of 4
An extract from one worked example. The figures are illustrative, not a customer’s result.
Check the numbers before you order stock
Describe the product and who buys it. Research and the full scored report cost nothing.
Or create an account first. Nothing you type here is lost when you sign up.