Amazon Product Research Framework: How Agencies Find Winning Products

Find winning products

Get expert product research before you launch

Our team evaluates demand, margins, and competition so you invest in products that scale.

Book a FREE 1:1 Strategy Call

Reading demand signals the right way

Demand validation is the first gate. If nobody is searching for what you plan to sell, nothing else matters. But "demand" on Amazon is more nuanced than a single search volume number. Here is what to look at and why each signal matters differently.

Search volume and keyword breadth

Start with your core keyword. Tools like Helium 10, Jungle Scout, and DataDive estimate monthly search volume on Amazon. A primary keyword with fewer than 3,000 monthly searches is a warning sign for most categories: there may not be enough organic traffic to sustain a new ASIN without heavy ad spend. But volume alone is misleading. A product with 15 relevant keywords averaging 2,000 searches each can outperform one with a single keyword at 30,000 if those secondary terms convert well.

Map the full keyword universe. Identify the top 20-30 terms buyers use, including long-tail variations and problem-aware searches. If the keyword cluster is deep, the demand is real. If it is thin, one or two terms carrying all the volume, you are exposed to ranking volatility. For a deeper breakdown of keyword mapping, see our Amazon keyword research guide.

Brand Analytics and Search Query Performance

If you already have Brand Registry (or a client does), Brand Analytics gives you something third-party tools cannot: actual Amazon click-share and conversion-share data tied to specific search terms. The Top Search Terms report shows which ASINs capture the most clicks for a given query. When a few ASINs dominate click share, the category is entrenched. When click share is fragmented across many ASINs, there is room for a new entrant.

The Search Query Performance API goes further. It shows impressions, clicks, cart adds, and purchases for queries your brand appears on. For existing sellers evaluating line extensions, this is the closest thing to a demand crystal ball. You can see exactly which queries drive purchases and which ones leak traffic to competitors. If you are pre-launch and lack this data, partner with someone who has category-level access.

Trend direction matters more than snapshot volume

A keyword at 10,000 monthly searches trending down 15% year over year is worse than one at 5,000 trending up 30%. Check Google Trends for the broader search interest curve. Cross-reference with Amazon-specific tools. Seasonal products are fine if you know the seasonality, but a product riding a fad that peaked last year is a money pit.

How high is the wall? Competition and review barriers

Demand without a realistic path to page one is just data. You need to evaluate what stands between you and organic visibility.

Pull the top 20 organic results for your primary keyword. Record: average review count, average star rating, number of Brand Registered sellers, presence of A+ Content and Brand Stories, and the ratio of FBA to FBM listings. Then look at the sponsored placements. Count how many sponsored slots appear on the first page and note whether the same brands occupy both organic and paid positions.

Here is what the numbers tell you:

• Average reviews above 1,000: High barrier. You will need a large review velocity strategy and significant ad budget to break in.
• Average reviews between 200-500: Moderate barrier. Achievable with a strong launch plan and differentiated listing.
• Average reviews below 200: Lower barrier, but also signals the category may be early-stage or low-volume.
• Star rating clusters at 3.5-4.0: Opportunity. Buyers are not fully satisfied. Read the negative reviews to find product improvement angles.
• Multiple Brand Registered sellers with premium A+ Content: These brands invest in creative and are harder to displace. Read our piece on why copycat selling is over to understand why matching these listings feature-for-feature will not work.

One overlooked signal: check if Amazon itself sells in the category. If Amazon Retail (sold by Amazon.com) occupies the Buy Box for the core search term, your organic and paid costs increase substantially. Amazon rarely loses the Buy Box on its own listings.

Margin and fee modeling: the math that kills bad ideas

A product can have strong demand and weak competition and still fail. If the unit economics do not work after Amazon takes its share, you are running a charity.

Build a per-unit margin model that includes every cost. Here is the structure we use:

1. Product cost (landed): Manufacturing, packaging, quality inspection, freight to Amazon's warehouse. Not just FOB price. Include tariffs and duties at current rates. With 2026 tariff structures still shifting, build two scenarios: current rate and a 10-15% increase.
2. Amazon referral fee: Typically 8-15% of sale price depending on category. Check the Seller Central fee schedule for your specific category.
3. FBA fulfillment fee: Based on product size and weight tier. Use Amazon's FBA Revenue Calculator with your actual product dimensions. Oversize products face dramatically higher fees.
4. Storage fees: Monthly and, if applicable, aged inventory surcharges. Factor in realistic sell-through rates, not best-case. See our FBA inventory guide for forecasting tips.
5. Advertising cost: Estimate your TACoS (Total Advertising Cost of Sales) at 15-20% for the first six months. Mature products may run 8-12%, but new launches rarely achieve that early. Understand how PPC economics work before assuming low ad costs.
6. Returns and defects: Category average return rates range from 3% (supplements) to 20%+ (apparel). Every return costs you the product, the FBA fee, and often a processing fee on top.

After all of this, your net margin should be at least 25-30%. Below 20%, you cannot absorb a bad month, a fee increase, or a competitor price war. Below 15%, walk away.

One common mistake: modeling margin at your target price without checking if that price is competitive. If every page-one product sells for $19.99 and you need $29.99 to make the math work, you do not have a viable product at that cost structure. Either reduce COGS or find a different product.

Where will you be different?

Differentiation is survival. Amazon's algorithm rewards products that convert. A listing that looks identical to ten others on the page will split clicks evenly and never build velocity.

Differentiation angles fall into a few categories:

• Functional improvement: Better materials, additional features, or solving a specific complaint found in competitor reviews. Example: a yoga mat with alignment lines when all competitors sell plain surfaces.
• Bundle or kit strategy: Combining the main product with a high-value accessory that competitors sell separately. This changes the perceived value and makes direct price comparison harder.
• Brand and visual positioning: Superior packaging, lifestyle photography, and A+ Content that signals premium quality. In commoditized categories, brand perception can justify a 20-30% price premium.
• Size or quantity variant: Offering a count or size that no one else stocks. If every competitor sells a 30-pack, a 60-pack with a better per-unit price can own that keyword variant.

The test for a real differentiation angle is simple: can a shopper see why your product is different within three seconds of seeing your main image and title? If you need the bullet points to explain it, it is too subtle. Main image and title carry over 80% of the click-through decision on a search results page. Invest in listing optimization early, not as an afterthought.

The go or no-go checklist

After completing the analysis above, run the opportunity through a binary checklist. Every item gets a pass or fail. One fail does not automatically kill the product, but two or more should.

1. Demand confirmed: Primary keyword cluster shows 10,000+ combined monthly searches with stable or growing trend direction.
2. Competition is beatable: Average reviews on page one are below 500, or a clear gap exists in product quality, branding, or listing execution.
3. Net margin above 25%: After all Amazon fees, advertising at 15-20% TACoS, returns, and landed COGS, per-unit profit meets or exceeds target.
4. Clear differentiation: You can articulate in one sentence what makes this product visibly different from the top five competitors.
5. Supply chain viable: Lead times, MOQs, and supplier reliability support a realistic launch timeline. You can afford to order enough inventory for 60-90 days of sales without cash flow risk.
6. Category risk acceptable: No gating, restricted brand issues, high regulatory burden, or IP risk that could trigger account suspension.

Score each item. If demand, margin, and differentiation all pass, you have a launch candidate. If two of those three fail, stop. The product is not ready, or it is not the right product.

What separates research from guessing

The framework above takes 10-20 hours per product opportunity when done properly. That time investment saves tens of thousands in failed inventory, wasted ad spend, and opportunity cost. Agencies run this process before every launch because the data is available to anyone willing to do the work. The gap between sellers who scale and sellers who churn is almost never "they found a secret product." It is that they validated harder and killed bad ideas earlier.

One final point: revisit your research every 90 days after launch. Demand shifts, new competitors enter, and Amazon's fee structure changes. The product that passed your checklist in Q1 needs re-evaluation by Q3. Treat product research as an ongoing discipline, not a one-time event.

Your next product decision should be backed by data, not intuition. Hyperzon's product research team runs this exact framework for DTC brands, identifying opportunities with real margin and defensible differentiation. Get a free Amazon audit.

Article was originally published on 02 October, 2026