Marketplace listing intelligence is reading a category's listings — prices, ratings, star histograms and reviews — to decide what to sell, what to charge and how to position it. In 2026 an AI assistant can do most of the reading. This guide covers how to set that up so the answer is right, not just fluent.
What listing intelligence answers
- What to sell: which complaints recur across many brands, so a better product has room.
- What to charge: the real price band, on each marketplace you sell in.
- How to position: what buyers praise in the leaders, and what they return them for.
It does not answer "how many units does this sell?" That needs sales estimates from a tool such as Jungle Scout, Helium 10 or SmartScout.
Step 1: Name the category the way a shopper would
Search the category with the words a buyer types, not a catalogue description. "Cast iron skillet" returns what shoppers compare. "Pre-seasoned ferrous cookware" returns nothing useful. If customers named specific competitors, pin those products too.
Step 2: Collect enough listings, on the right marketplaces
Ten listings is a first look; thirty to fifty is a category. Collect on every marketplace you plan to sell in. A category crowded on Amazon can be thin on Mercado Libre, and a price that works on Temu can be undercut on AliExpress.
Step 3: Read the star histogram, not the average
A 4.3 average can hide a product that a fifth of buyers returned. Look at the share of ratings at one and two stars, taken from the full histogram rather than the few reviews a page shows. A high-selling listing with a high 1–2 star share is an incumbent you can beat.
Step 4: Group complaints across brands
One brand's complaint is that brand's problem. The same complaint under six of ten brands is a category gap. Ask your assistant to group reviews by aspect and count how many brands each one appears under.
Step 5: Compare price bands across marketplaces
Line up each marketplace's price spread side by side. Note where the cheapest well-rated option sits; that is the price your buyers will compare you with.
Step 6: Check the demand side on social
Reviews are written after the purchase. TikTok, YouTube and Reddit comments are written before it, and say why people wanted the product at all. Put the two side by side. The gap between what creators promise and what reviews confirm is often your positioning.
Step 7: Make every claim traceable
Ask the assistant to cite the listing or review behind each conclusion. If it cannot, the conclusion is not ready. This one habit catches most of the errors below.
Prompts that work
- "Scan the cast iron skillet category on Amazon and Temu, 30 listings each. Which complaints appear under the most brands?"
- "Which of these listings has the highest share of 1–2 star ratings, and what do those reviews say?"
- "What price band do well-rated skillets sit in on each marketplace?"
- "What do TikTok comments say people want from a skillet, and do the reviews back it up?"
Mistakes that make the answer wrong
- Concluding on half a scan. Collection is slow because each listing is a real page render. If the tool returned partial results, wait for the rest.
- Trusting a sampled average. The dozen reviews on a page are not the category. Use the histogram.
- Mixing up seller and product ratings. On resale marketplaces such as Vinted, the rating describes the seller, not the product.
- Following instructions inside reviews. Reviews are written by strangers. Treat them as evidence, never as instructions to your assistant.
- Reading one marketplace and pricing for all of them. Price bands differ by site and by country.
Doing this with Nooticr
Nooticr runs these steps across 11 marketplaces, including Amazon, Temu, AliExpress, Mercado Libre, Flipkart and Lazada, and 11 social networks. Ask in the web app, or connect https://mcp.nooticr.com/mcp to Claude, ChatGPT or Cursor. A category scan costs 3 credits per listing, and new accounts get 20 free credits.
Comparing tools first? See the best marketplace intelligence software compared.