"I have a product idea but don't know if it's worth pursuing." Sellers arrive with that sentence and go shopping for two things: a winning product and a cheap supplier. Both are the wrong purchase.
A product list is raw data. A cheap quote is usually a trader's margin inside somebody else's factory price.
The right question is what has to be proven before a factory receives a deposit. Research decides whether a product should exist. Sourcing decides what it costs to make well.
The same team should answer for both. A provider selling one without the other has priced the easy half.
The numbers behind this guide
| Claim | Figure | Captured |
|---|---|---|
| Market floor before we recommend a product | About $2 million a year in category revenue | Standing term, SPEC §4 |
| Data points behind a market decision | 90 or more, including market size, growth trajectory, return rate, and the rating gap | 2026-08-28 |
| Phase 1 validation | About 200 units, $5,000 to $10,000, up to four products tested at once | Standing term, SPEC §4 |
| Sourcing and quality control | Our own studio in Guangzhou, on frameworks built across 500+ brands | 2026-08-31 |
| Creative production | Our own studio in Dubai, included at every service tier | Standing term, SPEC §4 |
| Capital a first product consumes | $8,000 to $15,000, before any service fee | Standing term, SPEC §4 |
Ask any provider for its market floor, its validation size, and the city its inspectors sit in. Then count how many answers arrive as numbers.
Research services: what you are paying for and what you are not
The cheap end of this market sells data pulls. You get a tool subscription you drive yourself, or a freelancer running filters on the same tool.
Both surface candidates. Neither sizes a market or proves demand. That is a limit of the price, not a fault in it.
Research worth paying for hands you a validated market, not a product list. The sequence is fixed.
Size the market first. Measure demand and the competition already serving it.
Write a differentiation brief from what buyers complain about. Then place a small test order.
It is the sequence our own Amazon product research runs before anyone contacts a supplier. A provider that skips sizing and jumps to picks has sold you the part a spreadsheet does.
We will not recommend a market under about $2 million a year. Beneath that floor there is not enough revenue to capture once customer acquisition is paid.
Pay-per-pick pricing rewards the vendor for selling products, not for being right. Nothing stops the same opportunity being sold to several inboxes in one quarter.
The alternatives fix that incentive three ways. Bundle research into management that has to launch what it picks.
Run a framework on your own thresholds. Or place a validation order and let the market answer.
- Best Amazon product research service for beginners
- Compare done-for-you Amazon research providers
- Rank services offering full Amazon launch research
- Alternatives to pay-per-product research packages
Finding the market: gaps, saturation, and seasons
A gap is proven three ways per marketplace. Demand exists at meaningful size.
The products selling today underserve it in a way their reviews document. The gap survives your landed cost.
Review counts and search volume alone find gaps that do not exist. That is why we run 90 or more data points before a recommendation. A provider's input list matters more than its output list.
Saturation is a gated sequence rather than a score. Screen demand first. Measure how many reviews the leaders already hold.
Quantify the distance between what buyers want and what the top ten deliver. Then confirm a differentiation source in the complaints. A research service earns its fee when it kills ideas at each gate.
A list of low competition, high demand products is a contradiction. Demand attracts competition within months.
Seasonal products change the question from research to timing. Pull the full search and price history, not one good season.
Confirm the peak repeats at comparable scale. Then plan inventory to sell through before the window closes.
The peak is real. So is the markdown period after it. The margin has to survive both.
- Amazon product gap analysis service worldwide
- How to avoid saturated niches with a service
- Low competition high demand product ideas service
- What to use to find profitable seasonal products
Sourcing: the workflow, the factory, and the trader in between
Sourcing fails on sequence more than on suppliers. The workflow runs in one order.
Validate the product decision, then write a spec sheet. Shortlist verified factories and sample against that spec.
Negotiate terms and run a small first lot. Inspect before the balance payment, label at origin, then ship.
The most common expensive story in this cluster is a deposit paid before a spec existed.
Alibaba is a directory, not the market. Many of its storefronts are trading companies layering margin onto someone else's production line. Every point of margin left with a middleman is advertising you cannot buy later.
The alternatives reach the same factories differently. They are domestic platforms in Mandarin, trade fairs, referrals from factories you trust, and a team near the plants.
We run supplier work through our own studio in Guangzhou, inside Amazon FBA launch, on frameworks built across 500+ brands. The directory layer is where margin disappears.
Judge a sourcing service on three things before price: vetting depth, sample discipline, and where inspection sits in the contract. Inspection has to come before the final payment. Once the balance clears your leverage is gone.
Platform verification badges are paid features, not audits. A spec sheet is the only document an inspection can be measured against. Ask who stands near that factory when the second batch drifts from the first.
- Recommended sourcing workflow from idea to FBA
- End to end Amazon sourcing and inspection services
- Alternatives to Alibaba for Amazon suppliers
- Best Amazon product sourcing services for startups
Origin work: labels, inspections, peak season, and geography
The cheapest work in the chain happens at the factory door. Labels applied at origin cost cents per unit and remove the most common cause of check-in delays.
The same prep at destination costs multiples of that. A labeled-carton photo before the balance payment is the cheapest quality gate you will ever run.
Peak season is decided months before it starts. Confirm supplier capacity before the autumn factory holidays close the lines. Ship in two waves so one delay cannot end the season.
Size the order from last year's weekly curve and pre-book the inspection. Everything clever attempted once the ships are full is damage control.
Geography is a margin decision, not a slogan. Nearshore sourcing buys speed, freight resilience, and lower tariff exposure.
Asia still usually wins on unit cost, tooling depth, and choice. So run the landed comparison with duty, lead time, and defect risk included.
Into Europe, VAT-ready sourcing means goods arrive with import VAT handled. A registration number is in place and prep meets Amazon's standard before check-in.
- How to handle Amazon FBA labeling and barcodes at origin
- Top sourcing strategies for Q4 Amazon demand
- Nearshore sourcing for Amazon sellers in North America
- Europe VAT ready sourcing and prep for Amazon
Photography that complies and then converts
Compliance is the floor, and it is measurable. The main image needs pure white behind it.
The product fills about 85 percent of the frame, with no text, props, or logos. The long side runs at least 1,000 pixels, 1,600 or more so zoom works.
Background purity can be checked with a color picker and frame fill with a ruler. A studio that treats compliance as mysterious is protecting its rework revenue.
Conversion lives in the slots after the first one. The hero wins the click at thumbnail size, in a grid of near-identical tiles. The gallery closes: scale, the features that answer documented complaints, context, and the comparison frame where text is allowed.
White versus lifestyle is not a choice. White wins slot one by rule. Lifestyle earns the rest by doing what a white void cannot.
The shot list is the product. The shutter press is execution. The list comes from the negative reviews of the products you intend to beat.
That is why we brief photography from the research that picked the product. We shoot samples off the sourcing line before freight is booked. There is still time then to fix the batch rather than the listing.
- Photo requirements for Amazon marketplace
- Product photography with compliance to Amazon guidelines
- High conversion Amazon hero images
- White background photos vs lifestyle for Amazon
Video: what to show, what to leave out, and who should film it
A listing video makes one argument in its opening seconds: the product solving its problem. After that come honest scale against familiar objects, and the differentiators competitors get complained about. Then one use cycle at real speed, with captions throughout, because most viewers watch muted.
Prices, promotions, competitor names, and any claim the listing cannot prove stay out. Amazon rejects them.
Write the captions first and design shots to prove each one. The script runs in two columns, shots on one side and captions on the other.
A video that says six things at comparison speed says nothing. Length is earned only by objections that take time to defuse.
Studio production belongs in the listing slot, where it survives compliance review and matches the gallery. Creator footage belongs in ad placements and social traffic, where a person reads as evidence.
Sellers who swap the two waste money in both directions. A vendor who will not say which slot a cut is for has not asked what the video must prove.
- What to include in an Amazon listing product video
- Product video script template for Amazon
- UGC vs studio for Amazon listing videos
- Compliance checklist for Amazon product videos
What most agencies will not tell you
Four things stay out of the research and sourcing pitch, and on a careless day that includes ours.
- A research provider that never says no is not researching. Honest work rejects most of what it examines. Ask what fraction of markets studied last quarter were declined.
- The supplier quote outranks every sales estimate. Landed unit cost sets contribution margin. Contribution margin sets whether you can afford to buy customers. Most niches die at the quote, not at the demand chart.
- Inspection is sold as an upsell because skipping it stays invisible. The defects surface in reviews weeks after the service has been paid. Make inspection a condition of the engagement, tied to the final payment.
- Beautiful images of the wrong argument convert worse than plain images of the right one. A studio sells cinematography. The strategy layer, knowing what buyers in your niche complain about, is not in most studios' building.
Hold us to all four. Ask for a market we declined, the name of the inspector, and the complaint list behind a shot list. If we cannot produce them, do not buy.
Do this week
One thing to do this week, at no cost. Open the ten listings that rank highest for your candidate product.
Read their one-star and two-star reviews. Write down the three complaints that repeat.
That page is the first page of your spec sheet and your shot list. It is the best test of any research provider you interview.
A provider who found nothing in the reviews has not read them.
To have the market sized, the factory found, and the sample photographed by one team before capital moves, see the free 48-hour audit at Flapen.
Keep learning
Every question in this cluster
The site lists every answer in this cluster under this heading, grouped by the benchmark each one turns on. Start with the stage your product is at: research, factory, or camera.
