How to Find Products to Sell on Amazon: A Research and Compliance Framework

Most first Amazon products fail before they’re ever listed. Someone picks a product because it looked hot on TikTok or a tool flagged it as “low competition”, skips the boring parts, and finds out three months later the product can’t legally ship or can’t turn a profit once fees and ads are subtracted. The framework below fixes the order of operations for how to find products to sell on Amazon: research the demand, then the money, then the legal exposure, in that sequence, not backwards.

What You Need Before You Start

You don’t need much to begin this process, but a few things will save you time. An active Seller Central account, Individual or Professional, gets you in the door, though Professional (at $39.99/month) unlocks Product Opportunity Explorer and removes the per-item fee that makes Individual plans expensive past a handful of sales. You don’t need a paid Amazon product research tool yet. The first five steps below run entirely on Amazon’s free tools plus a spreadsheet.

Budget one to three weeks for the research itself before you order a single sample, and plan for several more weeks after that for sample review and supplier back-and-forth. If you’re hoping to launch inventory in two weeks total, something in this process will get rushed, and it’s usually the compliance or trademark check, which is exactly the wrong place to cut corners.

Step 1: Choose Your Business Model First

Private label, wholesale, retail/online arbitrage, and dropshipping are not four flavors of the same business. They have different capital requirements, different margins, and different research processes, so picking one before you start hunting for products saves you from wasting time on the wrong kind of research.

Private label (building your own branded product) usually needs $3,000 to $10,000+ in starting capital and takes four to six months before the first sale, but margins typically land in the 20-40% range once you’re established. Wholesale (buying an existing brand’s inventory to resell) needs less upfront capital, often $2,000 to $5,000, but margins compress to 10-20% because you’re competing with other sellers of the identical product. Retail and online arbitrage can start with $500 to $2,000 and get you selling within weeks, but the supply is unpredictable. You can’t restock a clearance find.

We work almost exclusively with private label and established brands, so the rest of this guide leans that direction. If you’re doing wholesale or arbitrage, Steps 1 through 3 still apply. Steps 7 and 8 (compliance and trademark clearance) matter less if you’re reselling an existing, already-cleared product, though you still need to confirm your account is eligible to list it.

Step 2: Amazon Product Research Tools You Can Use for Free

Amazon’s Product Opportunity Explorer groups customer search terms into niches based on what people actually view and buy after searching, and it’s built into Seller Central at no extra cost for Professional accounts. Amazon states that new products launched from niches identified in Opportunity Explorer saw 2.5x higher sales potential in their first three months, based on 2025 internal data across US, EU, and Japan sellers. That’s a real signal worth checking before you spend a dollar on third-party software.

Access it under Growth, then Product Opportunity Explorer, in Seller Central. Search a category or customer need rather than a specific product name at first, since the tool needs enough search volume to group results into a real niche. Bathroom storage will surface more useful niche data than “under sink organizer with drawers.”

Best Sellers and Movers & Shakers are worth checking alongside it. Best Sellers shows current category leaders. Movers & Shakers shows what’s climbing fastest in the last 24 hours, which is a useful early signal but a noisy one. A product can spike on Movers & Shakers because of a single influencer post and fall back out within a week, so treat it as a lead worth investigating, not a conclusion.

Amazon’s search autocomplete is the most underused free tool in this list. Type a broad category term into the Amazon search bar and watch what it suggests. Those suggestions are pulled from real customer search behavior, and they’ll often surface a specific variant, size, or use case you wouldn’t have thought to search for on your own. Run the same term through Google’s search bar too. If Amazon and Google surface different modifiers for the same base term, that gap is worth a closer look.

Step 3: Set Numeric Filters, Then Adjust Them Per Category

Fixed thresholds don’t transfer cleanly across categories. A “good” Best Sellers Rank in kitchen gadgets means something different than in supplements. Still, you need a starting filter or you’ll drown in options while trying to find products to sell on Amazon. The ranges below are a blend of the thresholds seller research and research tools most commonly converge on, not a rule from Amazon itself, so use them as a first pass and adjust once you see what’s actually normal in your specific niche.

Treat these as a screen, not a scorecard. A product that fails one filter by a small margin but wins clearly on the others is still worth a closer look. A product that fails three of the five isn’t.

Step 4: Mine Reviews for What to Build Differently

This step turns “find a product” into “build a better product.” Most beginners skip it because reading 150 reviews sounds tedious. It’s also the single best hour you’ll spend in this whole process, and it’s the difference between launching a copy and launching something people actually have a reason to switch to.

Pull the 1-star, 2-star, and 3-star reviews on the top 5-8 competing listings. Skip the 5-star reviews for this exercise. They confirm the category has demand, but they won’t tell you what to fix. Log every complaint in a simple spreadsheet with a mention count next to it. Patterns show up fast. If “zipper breaks” appears 30 times across three competitors, that’s not one bad batch. That’s a design flaw the whole category shares, and it’s sitting there in plain sight for anyone willing to read past the star rating.

Your opportunity isn’t inventing something new. It’s fixing what customers already complain about in a product they’re already buying. A travel organizer with a reinforced zipper, better compartment sizing, and a washable lining isn’t a new product category. It’s the same category with the three most common complaints solved. That’s usually enough reason for a shopper to pick a 40-review listing over a 4,000-review one sitting at 3.9 stars.

Once you’ve identified what to fix, the same complaint data feeds directly into your listing copy and images.

Step 5: Cross-Check With a Paid Amazon FBA Product Research Tool

Amazon’s free tools are a strong starting point, but they don’t show you competitor pricing history, estimated revenue, or keyword-level search volume with the depth a paid tool provides. Tools like Helium 10, Jungle Scout, and Keepa each pull this data from slightly different sources and estimation methods, so numbers will vary between them. Use one to sanity-check what you found in Steps 2 through 4, not as your only source of truth.

Keepa is particularly useful here because it shows 90 to 365 days of price and sales rank history for a listing, which tells you whether demand is steady or whether you’re looking at a seasonal spike about to disappear. A product with a flat, stable BSR line for the last year is a much safer bet than one that spiked in November and has been sliding since.

Run one concrete check before moving on: pull the estimated monthly search volume for your top 2-3 candidate keywords. A niche where demand concentrates in a single keyword is fragile, since one algorithm change or a competitor’s ad spend can take a big share of it overnight. A niche where demand spreads across five or six related search terms is a steadier bet, even if each individual term looks smaller on its own.

Step 6: Run the Real Profit Math

Most “winning product” lists fall apart at this step, because a $30 selling price with a $5 product cost looks like a 500% markup on paper and is nowhere close to that once Amazon takes its share. Here’s what actually comes out of that $30 before you see a dollar of profit.

Amazon’s referral fee runs around 15% for most categories, so that’s $4.50 gone immediately. FBA fulfillment and storage fees depend on size and weight, but a small, light item typically runs $4-$6. Advertising costs vary widely, but a reasonable planning number for a new launch is 15-25% of your selling price, so figure another $4.50-$7.50. Then there’s landed cost, meaning product plus packaging plus freight plus duties, not just the factory quote. For a simple item that might run $5-$7 once everything is included, and sellers who only budget the factory price get an unpleasant surprise the first time a customs invoice arrives.

Here’s how that stacks up on an actual $30 product:

That’s workable. Drop the price to $18 with the same cost structure and the margin nearly disappears once fees and ads come out, even though the markup over factory cost still looks fine on a spreadsheet.

This is exactly why the $20-50 price band shows up as a starting filter across most seller research. Below that range, fixed costs like FBA fees eat a disproportionate share of the sale. Above it, return rates tend to climb and the buyer decision takes longer, both of which cost you conversion rate and, indirectly, ad efficiency.

Amazon’s Revenue Calculator will run these numbers for a specific product once you have real dimensions and weight. Run it before you place any inventory order, not after you’ve already committed to a supplier and a price point.

Landed cost is where most first-time importers underestimate their own math, because “landed cost” means more than product plus freight. If you’re importing, US Customs and Border Protection requires the correct HTS (Harmonized Tariff Schedule) classification for your product, and the duty rate attached to that classification can run anywhere from 0% to over 20% depending on the product category and country of origin. Country-of-origin marking on the product itself is also a CBP requirement, not an optional label. Get a customs broker or freight forwarder to confirm your product’s HTS code and duty rate before you finalize a price with your supplier, not after your first shipment gets held at the port for a classification you guessed at.

Step 7: Check Compliance Requirements Before You Order

Confirm the category doesn’t require testing, certification, or approval before you commit to inventory, not after the shipment is already on the water. A product that fails compliance after it’s paid for isn’t a design problem anymore. It’s a write-off.

The Consumer Product Safety Commission sets testing, certification, and labeling requirements for most consumer products sold in the US, and as of July 2026, importers of most regulated products must file certificate data electronically under CPSC’s updated process. Children’s products fall under CPSIA and typically need third-party lab testing for lead and phthalates before they can legally ship. Anything with a plug or battery usually needs UL or ETL certification. Neither of these shows up in a supplier’s price quote unless you specifically ask, and both can add weeks to your timeline if you find out about them after the order is placed rather than before.

Cosmetics and skincare carry their own layer under the FDA’s MoCRA requirements, which can involve facility registration and product listing depending on your business size and product type. Supplements, anything ingestible, and anything making a health claim carry even more scrutiny. If your product idea touches any of these categories and you don’t already understand the compliance requirements, a simpler, non-regulated product is the safer place to start.

Amazon also maintains a list of gated categories that require approval before you can list, including things like automotive parts, jewelry, and certain safety products. Check this inside Seller Central before sourcing in a category you haven’t sold in, since the list changes and gating sometimes applies to specific subcategories rather than the whole category.

Step 8: Clear the Product Legally Before You Commit

A great product idea that infringes on an existing trademark or design patent doesn’t stay great for long. This step takes an afternoon and can save you a five-figure inventory write-off.

Search the USPTO’s trademark database for your planned brand name before you order packaging, business cards, or a single unit. USPTO explicitly recommends this search before filing, because a name that’s “confusingly similar” to an existing mark can block your own registration even if you never intended to copy anyone.

Beyond the brand name, check whether the physical product design is covered by an existing design or utility patent. A quick Google Patents search for your product category will surface obvious conflicts. It won’t catch everything, and it isn’t a substitute for an actual attorney opinion on a product you’re about to invest heavily in, but it will catch the obvious problems before they become expensive ones.

Your supplier telling you a design “is fine to sell” is not legal clearance. Suppliers have no visibility into US trademark or patent filings and no liability if you get a cease-and-desist six months into your launch.

Clearing your brand name also opens the door to Amazon’s Brand Registry once your trademark is granted, which unlocks A+ Content, better counterfeit protection, and access to Amazon’s advertising formats reserved for registered brands. That’s a good reason to file early rather than treat trademark clearance as a box to check right before launch.

Step 9: Validate With Samples Before the Full Order

Order samples from two or three suppliers before committing to a production run, even if one quote looked clearly best on price. Compare material, stitching, finish, and packaging side by side. The cheapest factory quote is sometimes the most expensive mistake once you account for returns, refunds, and the reputation damage of shipping a product that doesn’t match its listing photos.

Ask each supplier if they can improve the product based on the complaints you logged in Step 4, not just replicate what already exists. That question alone tends to separate manufacturers who can support a real private-label relationship from ones who are only equipped to copy an existing design.

Packaging is worth testing at the sample stage too, not just the product itself. A product that solves every complaint from Step 4 but arrives in a flimsy box that dents in transit will still generate returns and 2-star reviews about “damaged on arrival,” which has nothing to do with the product you spent weeks validating.

Common Mistakes When Researching Amazon Products

Sellers go looking for the “best products to sell on Amazon” instead of checking whether a trending item’s demand has any staying power. By the time a product goes viral, dozens of other sellers have already placed their orders, and you’re launching into a market that’s about to be flooded.

They calculate margin off factory cost alone and skip referral fees, FBA fees, and advertising entirely, then wonder why a “50% margin” product barely breaks even. They skip the trademark search because it feels like an extra step with no immediate payoff, right up until it isn’t optional anymore. And they place a large first order to hit a supplier’s best price break before confirming actual sell-through, actual return rate, and actual advertising cost with a smaller test batch first.

I’ve also seen sellers treat a single BSR snapshot as proof of demand, when a product’s rank on any given day can swing based on a competitor’s stockout or a short-lived promotion. Pull the trend over 90 days, not the number you see today. And a smaller one: sellers assume a low review count always means low competition, without checking whether that’s because the product launched last month or because it’s genuinely a weak listing. Those are very different opportunities.

From Research to Launch

Using this framework to find products to sell on Amazon gets you halfway there. The other half is turning that validated product into a listing that actually converts, and that’s where a lot of otherwise well-researched products underperform. A product with a genuine complaint-driven improvement over its competitors still needs images, copy, and A+ Content that communicate that improvement clearly to a shopper scrolling search results in three seconds or less. Our Amazon listing design process picks up from exactly this point, turning the improvement you found in Step 4 into a listing that makes a shopper choose it over the competitor with 4,000 reviews.