A product test costs money twice: the ad spend, and the weeks you gave it. The cheapest way to waste both is to pick the product first and look for evidence afterwards. This guide does it the other way round. You start from products people are already buying, remove the ones you cannot make money on, and only then spend a small amount to find out the rest.
What “winning” means here
The word gets used for anything with a viral video. For the purpose of choosing what to sell, a winning product passes four checks at the same time:
- People are buying it now. Not last year, and not “it has a lot of views”. Units are being sold this week.
- There is still room. A few sellers doing well is a good sign. Dozens with the same video and a falling price is a closed door.
- The numbers leave room for ads. After the product, shipping and fees, one sale has to leave enough to pay for the ad that caused it.
- It sells itself in a few seconds. The problem and the result can be shown, not explained. That is what makes a cheap ad possible.
A product that passes three of four is usually a lesson, not a business. The steps below test them in the order that costs least.
The method
Start from sales, not from ideas
Open a list of products that are selling and work backwards. TikTok Shop is the most useful place to look, because every listing shows how many units it has sold. The winning product finder publishes ten such products every day with their units sold, price and rating, free and without an account. Read lists like this to learn what kind of product is moving, not to copy the first one you see.
Check the pace, not the total
Units sold is a running total since the listing went live. A product with 300,000 sold may have sold them two years ago. What matters is how many it sells per day now, which takes two readings some days apart. How to read TikTok Shop sales data explains the difference and the traps.
Look at the ads behind it
Find the ads that sell the product and note three things: how long the oldest one has been running, how many different advertisers there are, and what the ads actually show. An ad that has run for weeks is being paid for by someone who sees a return. Many new advertisers arriving at once with the same video is the sign of a crowd. The Meta Ad Library guide shows how to check this for free.
Do the margin sum before you fall for it
Take the price, subtract the product, the shipping and the fees. What is left is the most you can spend on ads per sale. A $35 product that lands at $9 with a 3% payment fee leaves $24.95, so its break-even ROAS is 1.40. A $19 product with the same costs leaves $9.43, and almost nothing converts cold traffic that cheaply. Break-even ROAS walks through the sum.
Check demand outside the app
Search the product’s plain name in Google Trends over twelve months. You are looking for the shape: steady, rising, seasonal, or one spike that has already come down. A seasonal product is fine if you are early for the season and a trap if you are late.
Check how crowded it is
Count the sellers and the advertisers, compare their prices and look at whether the pace is still rising. Six checks for saturation turns this into a yes, a no, or a “yes, with a different angle”.
Test small, with a rule for stopping
Write down the break-even cost per sale before the test starts, and decide in advance what ends it. A common rule: if an ad set has spent twice the break-even cost per sale without an order, stop it. If the first orders arrive under break-even, keep going and improve the ad. The rule matters more than its exact numbers, because it is made when you are calm.
What the evidence can and cannot tell you
Every number in this method is evidence about other sellers, not a promise about you. Units sold are counted on the listing, so they are as real as the platform’s own counter. Revenue figures are an estimate: price multiplied by units, which ignores discounts, bundles and refunds. An ad that has run for a month suggests it pays, but large brands also run ads that do not. Use each figure to remove bad options. The test at the end is the only step that measures your own offer.
The checklist
- It sold units this week, and you know roughly how many per day.
- The oldest ad for it has run for more than two weeks.
- Fewer than a handful of sellers use the same video and the same price.
- One sale leaves at least $20 before ads, or the price is high enough to carry a higher ad cost.
- Search interest is steady or rising, or the season is ahead of you.
- You can show what it does in the first three seconds of a video.
- You wrote down the stop rule before spending.
NicheHunter puts the first four on one screen: TikTok Shop sales with their pace, the ads on Facebook, TikTok and Pinterest behind a product, and the stores selling it. You can look around free and decide later whether Pro is worth it for you; pricing is here.



