Why many proxies are sold as-is, how that damages workflows, and what to ask for instead.
Many providers focus on raw capacity and marketing copy, instead of triage. They don't run thorough IP checks or bucket IPs into meaningful categories: clean / usable / risky / toxic. If they did, they could upsell genuinely clean IPs as premium, and assign the noisy ones to rotation-only tasks.
Across providers we tested (sample sizes between 100 and 10,000 IPs per provider), results show stark differences in usable proxy quality,
Apart from rotation logic and promised sticky times not being respected, we frequently observed as little as 10% of IPs being fit for sticky sessions in some pools, essentially a guarantee that long-lived sessions and multi-account workflows will fail. Industry averages for residential pools commonly fell in the 30–60% usable range, rotating and datacenter pools were often worse.
Important: a good IP today can become a bad IP tomorrow.
Almost no mainstream provider gives these guarantees, and that's why multi-accounting setups fail before anti-detect configs even matter.
Acquisition teams pile on, resellers inject cheap upstream IPs. The result: mixed pools with dongles, mobile tunnels, misconfigured NATs, recycled datacenter ranges, all sold under the same label.
Fallout: entire account fabrics and marketing campaigns become brittle or banned, sometimes retroactively.
Most users get detected because of bad proxy signals long before browser configs are a factor. Fix the proxy first. For guidance on browser setup, see: Anti-detect setup →
Our team (operating the reviewfixer account on BHW) has advised clients on proxy selection and anti-detect configs, produced post-mortems, and reported reproducible issues to providers. BHW Thread
Signals are combined, a "clean" IP passes all thresholds, "damaged" IP has actionable reasons.