Treat product research as a funnel: generate many candidate ideas cheaply, then screen them down using progressively more expensive checks. The mistake most new sellers make is doing this backwards — falling in love with one idea and then looking for reasons it will work, instead of generating a wide pool of candidates and letting the data eliminate most of them.

Where to generate candidate ideas

  • Marketplace browse/bestseller lists — a fast way to see what's already selling at volume in categories you understand. Bestseller rank is a lagging indicator (it tells you what's already won), so use it for category validation, not for finding an untapped gap.
  • Keyword and search-trend tools — surface what people are actively searching for, including rising queries that established sellers haven't caught up to yet. A search term with rising volume and few well-optimized listings against it is one of the highest-signal opportunities you can find.
  • Competitor gaps — read negative reviews on top listings in a category you understand; recurring complaints are often a specific, buildable product improvement. If five of the top ten listings in a category have the same 1-3 star complaint, that's not noise — it's a product brief.
  • Your own experience — categories where you have genuine buyer or industry knowledge give you a real research advantage over someone starting from zero. You already know the vocabulary buyers use, the quality issues that matter, and often a supplier relationship or two.
  • Adjacent-category logic — look at what sells well alongside a product you already understand (accessories, refills, bundles, a "pro" version) rather than starting a search from zero every time.
  • Supplier catalogs and trade shows — browsing a wholesale marketplace or supplier catalog by category can surface products you wouldn't have thought to search for, because you're seeing what's manufacturable rather than what's already being marketed.

How to screen candidates down

Run every candidate through three checks, roughly in this order because each is progressively more expensive to do properly:

  1. Demand check — is there real, current search/purchase demand? See Demand Research.
  2. Competitive check — can you realistically win against what's already there? See Competitive Research.
  3. Margin check — does the math work after realistic costs, not just landed cost? Run a rough pass with the Unit Economics Calculator, then a fuller pass with the Product Profitability Calculator once you have real supplier quotes.

Use the Product Opportunity Scorecard to score multiple candidates side by side rather than falling in love with the first idea that seems promising. Scoring several ideas on the same criteria at the same time also protects you from a subtle bias: an idea you've been thinking about for months will feel more validated than it is, simply because you've spent more time with it.

A funnel, not a single decision

Think in terms of ratios. A healthy research process might start with 20-30 raw candidate ideas, narrow to 5-8 after a quick demand/competition pass, and end with 1-2 that survive a full profitability and sourcing check. If your funnel only ever has one idea in it at a time, you don't have a research process — you have a hope.

A discipline worth adopting

Keep a running list of product ideas as you notice them (browsing marketplaces, reading reviews, seeing gaps in your own purchases) even when you're not actively looking — the best ideas often surface incidentally, and having a running list means you're screening from a real pool of candidates rather than trying to invent one idea on demand. A simple spreadsheet with columns for the idea, where you saw it, and a one-line reason it caught your attention is enough; you don't need special software for this stage.

Common mistakes at this stage

  • Stopping the funnel after generating ideas and skipping straight to sourcing the first one that "feels right."
  • Only searching in English/your home market when a product has different levels of competition on marketplace-specific storefronts elsewhere.
  • Treating a personal hobby interest as sufficient market validation on its own — passion in a category doesn't substitute for demand and margin data.
  • Ignoring operational fit — a product idea that's profitable on paper but requires cold-chain shipping, hazmat handling, or heavy regulatory review may not be a fit for your current operational capability even if the numbers work.

FAQ

How many product ideas should I evaluate before committing to one? There's no fixed number, but treat single-digit shortlists as too narrow if you're just starting out. Aim to seriously screen at least 5-10 candidates before committing capital to one.

Should I look for a product with zero competition? Usually not — near-zero competition often means near-zero validated demand. A moderately competitive category with a clear, buildable differentiation angle is usually a safer bet than an empty category with no proof anyone wants the product.

How long should product research take before I commit? It scales with how much capital you're about to commit. A low-cost, low-MOQ test order might justify a few hours of research; a large inventory commitment to an unfamiliar category deserves days to weeks of layered research across demand, competition, and margin.