Most "my survey accounts keep getting banned" stories I've heard don't start with the survey site. They start weeks earlier, at checkout — with a proxy purchase that had warning signs all over it. Here's how to spot a bad proxy before it costs you accounts. Seven flags, each with a concrete test you can run in minutes.

Flag 1: "Residential" IPs with datacenter ASNs

The most common scam in the proxy market: datacenter IPs marketed as residential. Survey sites check the ASN (the network owner) of your IP — a "residential" IP whose ASN belongs to a hosting company is an instant trust killer.

How to test: connect through the proxy, visit an IP info page (ipinfo.io, whoer.net), and look at the ASN/org field. If it says a hosting provider or cloud company instead of a consumer ISP, you're holding a datacenter IP with a residential label. Walk away.

Flag 2: No city-level targeting

Survey routing matches your IP's city against your stated demographics and timezone. A provider that only offers country-level targeting can't give you that match — and mismatched geo is one of the top screen-out triggers (see the setup guide).

How to test: before buying, check the provider's docs or dashboard demo: can you select a specific city or state? "Country only" is fine for casual browsing; it's not fine for surveys.

Flag 3: No sticky sessions

A survey is a session. If your IP rotates mid-survey, the panel sees one respondent become a different person halfway through — that's an automatic disqualification, and repeated occurrences flag the account. Any proxy sold for "account work" without sticky sessions is disqualified for surveys.

How to test: ask support directly: "Can I hold the same IP for 30+ minutes?" If the answer is vague, or sticky costs extra on top of an already-premium rate, keep shopping. Both providers I recommend in the comparison support sticky sessions.

Flag 4: No trial and no refund

A proxy provider that won't let you test before paying is asking you to gamble. Quality varies enormously between providers and even between their own pools — the only way to know is to run their IPs against your actual survey sites.

How to test: this one's binary — trial or money-back guarantee exists, or it doesn't. If neither exists, your "test" costs the full minimum top-up. Some good providers have no trial (Floppydata's minimum is $19, for example) — that's acceptable only when everything else checks out and independent reviews confirm quality.

Flag 5: Exit IPs on public blocklists

Some cheap pools are full of IPs that have been abused before — spam, botting, previous survey fraud. Survey panels check blocklists and fraud scores. You inherit the IP's reputation the moment you connect.

How to test: during the trial, run a few exit IPs through a fraud-score checker (scamalytics-style tools, or whoer.net's blacklist section). A couple of flagged IPs in a large pool is normal; a pattern of them is not.

Flag 6: "Unlimited residential" for a few dollars a month

Real residential bandwidth costs real money — providers pay the device owners. Anyone selling "unlimited residential proxies" for $3/month is selling you one of three things: datacenter IPs, a heavily shared pool where hundreds of users rotate through the same abused IPs, or nothing at all. The economics don't work otherwise.

How to test: compare against the verified market floor. Legitimate residential starts around $0.49–$1/GB in 2026 (see the cost math). If someone's "unlimited" price implies $0.01/GB at any plausible usage, it's not what the label says.

Flag 7: No usage dashboard

If you can't see per-GB usage, you can't verify what you're paying for, spot a leaking configuration (like a profile burning bandwidth on auto-playing content), or confirm the provider's accounting matches reality. No dashboard also usually means no session control and no targeting options — it's a proxy-shaped black box.

How to test: look at the dashboard screenshots on the provider's site before buying. You want to see: bandwidth usage, active sessions, targeting controls, and session duration. If the "dashboard" is just a buy button, pass.

The 15-minute pre-purchase routine

Put it together: (1) confirm city targeting and sticky sessions in the docs, (2) start the trial, (3) check 3–5 exit IPs for ASN type and blocklist status, (4) run the proxy against 2–3 of your actual survey sites for a day, (5) check the dashboard shows sane usage numbers. Fifteen minutes of work, and you'll never again wonder whether the proxy was the reason accounts died.

One more thing: even the best proxy can't save a bad setup. IP is one signal among several — fingerprint isolation, timezone matching, and human behavior matter just as much. The full picture is in the complete survey stack guide.

Question: which of these flags burned you before you knew to check? The "unlimited residential" one gets almost everyone once — share your story so beginners don't repeat it.