Route data is the part people forget
The most common failure in brand monitoring collections: a folder of screenshots with no record of which market or exit route produced them. Two weeks later someone asks which region showed the listing and the answer is a shrug.
The row beside the image matters as much as the image. URL and marketplace are obvious fields. Seller name, product name, claimed brand, and market are review fields. Timestamp, screenshot hash, proxy route, HTTP status code, provider byte count, and final human verdict are the fields that make the record usable after the page changes or the seller deletes the listing.
One exit gives you one view
A public listing can appear from Spain and return a 404 from Germany. The same listing can carry a different price or a different seller name depending on the region the request comes from. Datacenter exits work on open pages, but many marketplaces filter or modify content based on the ASN of the request.
Rotating residential proxies solve this because they place your request in the same IP range a buyer in that market uses. On the Proxynade pool at http://proxynade.net:2555, the country token in the expanded username sets the exit region. For a US marketplace check the username looks like rt97db6958d9-plan-volume-country-us-lifetime-10: base user, plan token, country code, and a rotation window in minutes.
The provider meter counts more than your crawler does
Your crawler counts saved rows, matched listings, and screenshots. The proxy meter counts redirects, images, thumbnails, scripts, block pages, failed seller pages, and discarded results. That gap is why app-level bandwidth estimates are usually low, and why the number on the bill is larger.
Screenshot mode makes the gap bigger. A text-only discovery pass can block images and stay light. A visual evidence pass may need galleries, badges, and product image assets loaded so the screenshot is worth keeping. Those are different capture modes and should have separate settings.
The dashboard usage logs on Proxynade show host, outcome, latency, and byte totals per request. Export them as CSV and join against your crawler output. The difference is what you paid for but did not keep.
Plan selection by check type
| Check type | Plan | Rate | Reason |
|---|---|---|---|
| Broad public discovery | Volume Residential | $0.89/GB | Failed assumptions stay cheap while you tune the run. |
| Marketplace with ASN filtering | Premium Residential | $5.00/GB | Move only the slice that keeps wasting retries, not the whole run. |
| Open pages, no filtering | Datacenter | $0.65/GB | Faster per request; use where ASN is not a factor. |
| Stable public route needed | Static ISP | Pay-per-IP | Consistent exit when rotation would disrupt a multi-step check. |
Start broad checks on Volume Residential. If one marketplace consistently blocks that route and forces retries, move only that marketplace slice to Premium Residential. Do not move the entire run on the basis of one site's behavior.
Seller profile checks need sticky sessions
Broad discovery wants rotation. Seller profile tracking wants the opposite: hold the same exit while reading the listing page, product set, feedback page, and seller metadata for a single seller. Swapping the exit mid-seller risks getting fragmented data from two different regional views stitched into one record.
The lifetime-<minutes> token in the username controls the sticky window. A seller pass that takes around ten minutes would use lifetime-15 with a buffer. Once the seller pass is complete, rotate before the next seller or market.
Scope stops at public pages
Brand protection monitoring on public marketplaces — listings, seller pages, pricing, regional availability — is the scope of this post. If the workflow needs login, checkout, private messages, or account actions, that is a permissions problem before it is a proxy problem. The Proxynade AUP prohibits unauthorized account creation and ban evasion.
Brand protection monitoring FAQ
Why does brand monitoring need residential proxies rather than datacenter? Marketplaces often show different content or block datacenter ASNs outright. A residential exit gives you the same view a buyer in that market sees.
What fields should every evidence record include? URL, seller, marketplace, timestamp, screenshot hash, proxy route, HTTP status code, provider byte count, and final human verdict.
Why does the proxy meter show more bytes than my crawler logged? The proxy counts redirects, images, thumbnails, scripts, block pages, and failed seller pages. Your crawler counts only the rows it kept.
When should I use sticky sessions during seller profile checks? Hold the same exit while reading the listing page, product set, feedback page, and seller metadata for a single seller. Rotate before moving to the next seller or market.
What proxy plan fits broad discovery versus targeted evidence collection? Start broad checks on Volume Residential at $0.89/GB. Move a specific marketplace slice to Premium Residential at $5.00/GB if that route keeps wasting retries. Datacenter works on open pages with no ASN filtering.
Collection checklist
- Record proxy route, exit country, and status code alongside every screenshot.
- Separate discovery passes (rotate freely) from seller profile passes (sticky session).
- Block images during text-only discovery; enable them only for evidence screenshots.
- Export usage logs from the dashboard and compare against crawler output.
- Stay on public pages; stop before any action that requires account credentials.