How to Validate a Product Idea Before You Spend Money on Stock

To validate a product idea, get strangers to do something that costs them money or effort before you buy much stock. The strongest proof is a paid pre-order or a sale, followed by an email signup from someone who found you cold. Compliments, likes and "I'd definitely buy that" count for very little. You can usually run the whole process for a few hundred dollars or less, and it should end in a clear decision: go, change the idea, or drop it.

The tests below run from cheapest to most convincing. Do them in order, and stop as soon as an idea fails. You should already have one product shortlisted. If you don't, start with how to find products to sell online.

What counts as evidence

Not all signals are equal. Here's how I'd rank them:

Keep your eye on the word "strangers." Your friends, family and followers are useful for feedback, but they aren't a test of demand.

Before you start, write down what would count as a pass for each test. If you set the target after seeing the results, you'll talk yourself into almost anything.

How to validate a product idea in five stages

Each stage costs a little more than the last, so a bad idea fails cheaply.

The costs and times above are my estimates for a simple physical product. Yours will vary with the supplier and the country.

Stage 1: Do the desk research

Start with a check you can do from your sofa:

• Use Google Trends to confirm interest isn't collapsing.
• Find at least three sellers who appear to be selling the same type of product.
• Read their negative reviews and list the repeated complaints.
• Work out the profit on one order using real quotes.

The last point catches many ideas. If a $30 product leaves $13.83 before marketing, you can afford some ad spend. If it leaves $4, no test will save it. Our profit margin guide shows how to get this number right.

Stage 2: Talk to real buyers

Find five to ten people in your target group and ask about their past behaviour, not your idea. People are poor at predicting what they'll do, but they can usually tell you what they've done.

Good questions:

• When did this problem last come up?
• What did you do about it?
• What have you bought to fix it? How much did you pay?
• What was wrong with it?
• Where did you look when you were choosing?

Bad questions:

• Would you buy this?
• Do you think this is a good idea?
• How much would you pay for this?

Don't mention your product until the end, if at all. Listen for the words people use, because they'll become your ad copy and product description. If nobody has ever tried to solve the problem, even with a workaround, then it may not hurt enough to spend money on.

Reddit, Facebook groups and niche forums are good places to find people, but follow each group's rules. Don't turn research into a disguised promotion. It's fine to say you're researching a product and ask if anyone will talk to you.

Stage 3: Get a sample in your hands

Buy a sample before you spend on marketing. Better still, order from two or three suppliers and compare them. Then use the product the way a customer would. Wash it, drop it, post it to yourself and leave it in the sun if that's what buyers will do.

You're checking three things:

1. Does it match the supplier's description and photos?
2. Would you be comfortable with your name on it?
3. Does it survive shipping in the packaging you plan to use?

This stage also gives you real photos and video for the next test. Avoid building your test on a supplier's stock images, since those same images are likely used by other sellers.

Stage 4: Run a landing page and ad test

This is the first test that uses people who don't know you. Build one simple page with:

• A clear headline stating who the product is for and what it does
• Real photos of your sample
• The actual price
• One action, such as "Join the waitlist" or "Pre-order now"

Showing the price matters. A free signup with no price tells you people are curious. A signup after seeing "$30" tells you far more.

Then send traffic from people who match your buyer. A small paid test works well because it reaches strangers quickly, and you control who sees it. Search ads catch people already looking, and social ads work better for products people didn't know they wanted. If you're going to do this, set your budget, your limit and your targets first.

Here's a worked example. The ad cost per click and the results are made up, so use your own numbers:

So the test has to bring in 8 orders to pay for itself. Seven orders would earn $96.81, and eight would earn $110.64.

You don't need to break even on your first test, though. A new page has no reviews and a rough first draft of its copy, and it'll improve. Say you got 3 orders from those 200 clicks. That's a 1.5% conversion rate and $33.33 of ad spend per order, or more than double the $13.83 you can afford. That isn't an automatic failure, but you'd need a better price, a better page or cheaper traffic before scaling.

A sample of 200 clicks is small, so don't read too much into small differences. It can show you the difference between "nothing happened" and "people clearly want this", but it can't tell you exactly what your long-term conversion rate will be.

Here's how I'd read the result:

Change only one major thing at a time, and give yourself two or three rounds at most. Past that, you're usually trying to rescue an idea you like.

Stage 5: Take pre-orders or sell a small batch

This is the real test, because people are handing over money. There are two ways to do it.

Pre-orders work when you can't order until you've proved demand. They suit products with a supplier minimum. For example, if a supplier needs a minimum of 100 units at $9 each, that's $900. At $30 each, 30 pre-orders would bring in $900 before fees, which would cover the product cost of that first order.

Be honest with pre-order customers. State clearly that the item ships later, give a realistic date, and offer an easy refund. If you sell to customers in the US, the FTC's mail order rule says you need a reasonable basis to ship within the time you state, or within 30 days if you state no time. If you can't ship on time, you have to tell customers and give them the option to cancel for a full refund. Other countries have their own rules, so check before you take money.

Never charge for something you have no real plan to deliver, and don't keep money from a test that doesn't go ahead. If your pre-order target isn't met, refund everybody quickly. Remember that some payment providers keep their original fee when you refund. Stripe says its processing fees from the original transaction aren't returned, as of November 2026, so count those as part of your test cost.

A small batch works if your supplier accepts a low minimum, or if you can make the product yourself. Pay a higher unit price and accept it. An order of 20 to 50 units at a worse price is often cheaper overall than being stuck with 500 units nobody wants. This is a good use of a small starting budget. If you don't want to hold stock at all, dropshipping can be used for a short test, but see dropshipping vs holding inventory for the quality and delivery problems it brings.

Your first buyers are also your best source of information. Contact each of them after delivery and ask why they bought, what nearly stopped them and what they'd change. Then watch your return rate and complaints, since a product that sells but comes back isn't validated.

Making the go, change or drop decision

At the end of the process, sit down with the numbers you wrote down at the start, and pick one of three outcomes.

Go if strangers paid or clearly tried to, your margin still works with the real cost of winning a customer, and early buyers are happy.

Change if there's interest but something is off. Maybe people want it in a different size, for a different use, or at a different price. Run one more test with that change.

Drop if strangers don't care after a fair test. This is a good outcome. Finding out for $200 is much better than finding out with $3,000 of stock in your garage.

Mistakes that make validation worthless

• Testing on your own audience. Your followers aren't your market, and your mother is definitely not your market.
• Hiding the price. Interest without a price doesn't tell you much about demand.
• Counting vanity numbers. Views and likes don't pay your supplier.
• Moving the target. If you said 5% and got 2%, you missed it. Don't decide that 2% was good all along.
• Testing a different product from the one you'd sell. A polished render can sell an item your supplier can't produce at that quality.
• Skipping the margin check. A test can show that people want something and still leave you with no business, if the cost of getting each order is more than you make on it.

What to do next

Write your pass targets for each stage today, before you've seen any results. Then book a week for your buyer conversations and order a sample.

Once an idea passes, you can start building with far less risk. How to start an online store covers the setup, and our guide to product descriptions will help you turn what you learned from buyers into a page that sells. Keep the email addresses you collected too, because those people are your first launch list. Email marketing for small online stores explains what to do with them.