The way to find a product to sell online is to run a filter, not to wait for an idea. You generate a wide list from places where real demand is already visible, cut it against fixed written criteria, then test the two or three survivors with a small amount of money before you commit to inventory. Inspiration is the cheap part. Filtering is the skill.
Most people do this backwards. They fall for a product first, then look for reasons it will work. This article gives you the order that stops that happening, along with the specific numbers you should be writing down at each stage.
Step 1: How to Find a Product to Sell Online Starts With Criteria
Set your filters before you start browsing, because every product looks good when you are excited about it. Put these on one page and do not move them mid-search.
- Price band. Decide the retail price range you will work inside. Low-priced items need volume to cover ad and fulfilment costs; high-priced items need more trust before a stranger buys.
- Margin floor. Pick the minimum net margin you will accept after product cost, shipping, marketplace or payment fees, and returns. Write the number down.
- Size and weight ceiling. Shipping cost scales with both. A hard ceiling removes a lot of bad ideas instantly.
- Fragility and perishability. Glass, liquids, and anything with a shelf life add breakage, packaging and compliance cost.
- Regulatory load. Supplements, cosmetics, children’s products, electronics with batteries, and anything medical-adjacent carry rules you must research before, not after.
- Seasonality tolerance. A product that only sells in December is a business you run one month a year.
These six lines are your gate. Anything that fails one of them is out, no matter how appealing it looks.
Step 2: Generate a Wide List From Places Demand Is Already Visible
Now go wide. The goal here is quantity — aim for forty or fifty candidates before you cut anything.
Good sources of raw ideas:
- Marketplace bestseller and movers-and-shakers listings. These show what is already selling, by category, updated constantly.
- Marketplace search suggestions. Type a category word and read the autocomplete. Those are real queries real people typed.
- Review sections of existing products. Three-star reviews are the richest ground in e-commerce. People who mostly liked something will tell you exactly what was wrong with it.
- Niche community threads. Hobby forums, subreddits and Facebook groups where people ask “what do you use for X?” give you products with a built-in audience.
- Your own last twelve months of purchases. You have domain knowledge somewhere. Use it.
- Supplier catalogues and trade directories. Browsing what manufacturers already produce shows you what is realistically available to make.
Put everything in one spreadsheet with a column for where you found it. Do not evaluate yet. Evaluating while collecting narrows your list too early.
Step 3: Cut the List Against Your Criteria
Now apply Step 1 ruthlessly. Most of your list will die here, which is the point.
For each surviving candidate, add these columns:
| Column | What to record | Why it matters |
|---|---|---|
| Retail price | Current typical selling price | Sets your margin ceiling |
| Landed cost | Unit cost plus freight, duty and per-unit packaging | The real cost, not the quoted one |
| Fees | Marketplace referral, fulfilment and payment fees | Often the difference between viable and not |
| Net margin | Price minus landed cost minus fees | Compare against your floor |
| Competing sellers | How many are selling something near-identical | Signals how hard entry will be |
| Review depth of top sellers | Review counts on page-one results | Deep review moats are expensive to climb |
| Differentiation idea | One sentence on what you would change | No answer here means no reason to buy from you |
That last column is the one that separates a product idea from a business idea. If you cannot write a sentence describing what makes your version worth choosing, you are planning to compete on price alone.
Step 4: Read the Complaints and Find the Gap
This is the step most people skip, and it is where the actual opportunity sits. Take your top five candidates and read fifty reviews of the leading products in each — not the five-stars, the two- and three-stars.
Tally the complaints. You are looking for the same problem appearing again and again: the strap breaks, the instructions are useless, it does not fit the standard size, the battery life is overstated, the packaging arrives crushed. A repeated, specific, fixable complaint is a product brief written for you by the market for free.
Then check whether the complaint is fixable at your scale. Some are: better materials, a different size, an extra accessory in the box, clearer instructions, a bundle that saves a second purchase. Some are not: a fundamental design limit, or a problem the category’s price point makes unavoidable. Keep the fixable ones.
Step 5: Check Demand Durability, Not Just Demand
A spike is not a market. Before you commit, test whether interest is stable.
Pull the search trend for your main product terms over at least the last two to three years using Google Trends, which is free and public. You are looking for one of three shapes: steady, seasonal-but-repeating, or a spike that has already peaked. Steady and seasonal are workable. A single spike with a declining tail is a product you will buy inventory for just as demand leaves.
Also sanity check how the product is bought. Is it a one-time purchase, a replacement cycle, or a consumable? Consumables and replacement-cycle products give you repeat revenue from the same customer, which changes what you can afford to spend acquiring that customer in the first place.
Step 6: Validate With the Smallest Possible Spend
Do not place a container order to test an idea. Validate in the cheapest form available for your model:
- Order samples. Three suppliers, one unit each. Handle the product. A surprising number of ideas end here.
- List a small quantity. A handful of units at real retail price tells you more than any amount of analysis. You learn the listing, the photography, the questions buyers ask and the return reasons.
- Run a tiny traffic test. A small ad budget to a real product page measures whether people click and add to cart, not whether they say they would.
- Talk to ten people in the niche. Not “would you buy this” — ask what they currently use and what annoys them about it. Then check whether your version answers that.
Write down in advance what result would make you stop. A pre-committed kill condition is the only reliable protection against sunk-cost thinking.
How to Find a Product to Sell Online: Putting It Together
Finding a product to sell online is a filtering process you can repeat as often as you like. Set criteria, go wide, cut hard, read the complaints, check durability, then validate small. Each round makes you faster at it, and the spreadsheet you build becomes a permanent asset.
Common Mistakes Worth Avoiding
Picking a product with no differentiation idea. If your only plan is to list the same item at a lower price, you have bought a race to the bottom.
Ignoring landed cost. The quoted unit price is not the cost. Freight, duty, packaging, inbound shipping and returns all land on the same unit.
Choosing a category with heavy regulation before you understand the rules. Research the requirements first; some categories need documentation you cannot get as a new seller.
Confusing your enthusiasm with market demand. Your interest helps you persist. It does not create buyers.
If you want the criteria, scoring columns and validation checks already laid out, work through the Amazon FBA Product Validation Kit on CourseVu and run your shortlist through it before you spend anything on inventory.