What marketplace research looks like
Marketplaces such as eBay, Walmart Marketplace, Etsy, Mercado Libre, Allegro and regional platforms are where many brands and sellers compete. Typical research tasks include:
- Competitor pricing and how it changes over time
- Seller counts and buy-box style competition for specific products
- Best-seller and trending lists by category and country
- Listing quality of competitors: titles, images, shipping terms
- Unauthorised sellers and counterfeits of your own brand
- Market sizing for a new category or country
Marketplaces protect their data and personalise results, so repeated automated checks from one IP don't last long. Geonode proxies let you spread checks across many IPs and see each marketplace from the right country.
Which Geonode product to use
| Marketplace behaviour | Geonode product |
|---|---|
| Light protection, public listing pages | Rotating datacenter (test first) |
| Bot protection on search and listing pages | Rotating residential |
| Region-specific results and shipping | Residential with country, state or city targeting |
| A few listings needing a browser and little maintenance | Scraper API |
Most large marketplaces need residential traffic for search and category pages. Some individual listing pages are more tolerant; test each page type separately.
Seeing the marketplace from the right place
Marketplaces vary search results, shipping options and sometimes prices by location. For research that reflects what buyers in a market see:
- Use an exit in the market's country, and the marketplace's local domain.
- Set delivery location where the marketplace offers it, inside a sticky session.
- Match language and currency settings to the market.
- Repeat checks from a few fresh sessions; personalised ranking means one session isn't representative.
Session settings
- Rotate per request for independent listing pages.
- Use sticky sessions of a few minutes for search results with pagination, and for any flow that sets location cookies.
- Space out requests. Marketplaces watch for rapid sequential searches.
Efficient collection
- Search pages first, listings second. Collect the list of relevant listings from search or category pages, then fetch only the listings you need.
- Prefer HTML or JSON over full browser loads.
- Re-check listings on a schedule that matches how often they change.
- Store raw responses briefly so you can re-parse without re-fetching when a parser breaks.
Residential traffic is billed per GB, so these habits keep costs low. Geonode's residential bandwidth doesn't expire, which helps when research is project-based rather than continuous.
Proxy tip: Before scaling, run a short pilot on your real targets with Geonode and record success rate, latency and GB per result. Our Geonode for price monitoring explains which Geonode product to start with.
Using the data well
- Normalise prices including shipping, since marketplaces show them differently.
- Track sellers by ID, not name, since names change.
- Separate sponsored placements from organic results in search data.
- Record country, session and timestamp with every observation.
Brand protection on marketplaces
If you're looking for counterfeits or unauthorised sellers, collect evidence carefully: screenshots, URLs, seller IDs, timestamps and the exit location. Many marketplaces have formal brand protection or rights-owner programmes; use your data to file reports through those channels. See Geonode for brand protection.
Staying within the rules
Marketplaces' terms generally restrict automated collection, and some offer official APIs for sellers and partners. Use official APIs for your own seller data, keep public data collection modest and polite, avoid personal data about buyers and sellers beyond business identifiers, and get legal advice for large commercial projects. See ethical web scraping.
Common problems
- Different results every run: personalisation and sponsored placements. Use fresh sessions and repeat checks.
- Shipping or price mismatches: location not set in the session. Set it explicitly.
- Search pages blocked, listings fine: search is more protected. Lower search rates and use residential for search only.
- Layout changes break parsers: marketplaces test layouts constantly. Log unparsed pages and alert on parse failures.
A sample tracking plan for a brand
A small consumer electronics brand sells on three marketplaces in two countries and wants weekly visibility.
| Item | Detail |
|---|---|
| Search terms | 15 category terms per marketplace and country, in the local language |
| Search depth | First two pages of results |
| Listings tracked | The brand's own listings plus the top 20 competitor listings per term |
| Fields | Position, price including shipping, seller ID, rating, review count, sponsored flag |
| Frequency | Weekly for search positions, twice weekly for prices |
| Proxy setup | Geonode residential exits per country; short sticky sessions for paginated search |
From this, the brand can see share of first-page placements, price position against competitors, and new competitors entering its categories. It also spots unauthorised sellers on its own listings, which feeds into brand protection work.
Common analysis mistakes
- Mixing sponsored and organic results inflates or hides real positions.
- Comparing prices without shipping misleads in marketplaces where shipping varies widely.
- Single snapshots: marketplace rankings fluctuate daily; trends matter more than any one day.
- Ignoring variants: a listing's price can change with the selected size or colour.
Avoid these and marketplace data becomes a reliable input to pricing, advertising and assortment decisions.
FAQ
Do marketplaces need residential proxies?
Search and category pages on large marketplaces usually do. Some listing pages work with datacenter traffic. Test each page type.
How do I research a marketplace in another country?
Use Geonode exits in that country, the marketplace's local domain, and local language, currency and delivery settings.
Can I track a competitor's sales?
Marketplaces don't publish sales directly. Researchers estimate demand from signals such as stock changes, review counts and ranking, which you can track over time from public pages.
What's the cheapest way to start?
A small, non-expiring residential balance, HTML-only fetching, and a short list of products to track. See cheap proxies.
Editorial note: Proxy providers change their pools, features and prices often, so treat any figures here as typical ranges at the time of writing and confirm current terms with the provider before you buy.