The Residential Proxy Audit: Why a 10-Million-IP Pool Still Gets Blocked (With an Honest Thordata vs. Decodo Comparison)

IP pool size is the most-quoted and least-predictive stat in the proxy market. What actually determines whether your traffic gets through: IP diversity across ASNs, geo accuracy at city level, session control, protocol support, and how the pool is sourced. This is a five-point audit you can run on any provider — followed by a fair, number-for-number comparison of Thordata and Decodo on exactly those points.

“Over 100 million IPs” reads well on a homepage. Then your scraper gets a 403 from a mid-tier retail site at 3 a.m., and the pool size number explains nothing.

The Five-Point Residential Proxy Audit

Run this audit against any provider — including the two compared below — before committing traffic.

1. ASN diversity. Anti-bot systems score IP reputation at the network level, not just the address level. A pool concentrated in a handful of ASNs burns collectively: once one subnet is flagged, every IP in it inherits the suspicion. Ask how many distinct autonomous systems the pool spans, and watch for answers that only quote address counts.

2. Geo accuracy, verified. “Covers 195 countries” says nothing about whether a request targeted at Osaka arrives from an Osaka IP. City-level and ASN-level targeting are table stakes; the test is whether the geo you request is the geo you get. Mismatched geos break regional price comparison, localized content testing, and ad verification simultaneously.

3. Session control. Rotating IP-per-request is right for high-volume single-page fetches. But logins, carts, and multi-step flows need the same IP for minutes at a time. Look for sticky sessions with configurable duration — Thordata supports sessions up to 30–90 minutes — and check whether session creation costs extra (on well-designed platforms, it doesn’t).

4. Protocol coverage. HTTP and HTTPS are universal; SOCKS5 matters for tooling that tunnels raw TCP. If a provider offers HTTP/HTTPS only, some of your stack will need a workaround.

5. Sourcing and ethics. How the pool is built affects both your compliance posture and long-term pool health. Thordata describes its residential pool as ethically sourced through an opt-in SDK model — worth asking any provider the equivalent question, because “residential” without a sourcing answer is a risk you’re quietly holding.

Running the Audit as Code

A practical audit is 30 lines. Pin a geo, hold a session, and verify what comes back:

import requests

# Grab the exact endpoint string from your provider's dashboard
# (Thordata ships an endpoint generator for location-pinned endpoints)
PROXY = {
    "http": "http://USERNAME-session-abc123-geo-jp-osaka:PASSWORD@gateway:PORT",
}

def audit_session(url: str):
    r = requests.get(url, proxies=PROXY, timeout=30)
    return {
        "status": r.status_code,
        "exit_ip": r.json().get("ip"),          # does it match the geo you requested?
        "asn": r.json().get("asn"),             # is the network residential?
    }

# Repeat the same session ID to test stickiness
print(audit_session("https://api.ip-check.example/json"))
print(audit_session("https://api.ip-check.example/json"))  # same session → same exit IP?

Run this across your ten most valuable target domains, at the volume you actually expect, and count delivered results per 100 attempts. That single ratio — not any marketing number — is your provider’s real grade.

The Numbers: Thordata vs. Decodo, Point by Point

Both companies are legitimate, established proxy providers. Decodo (formerly Smartproxy) built its reputation on approachable, consumer-friendly proxy products; Thordata built its platform around result-priced scraping with a large underlying pool. Here’s the published pricing on the residential side:

Where Decodo genuinely wins: entry simplicity. Its subscription tiers, 3-day trial, and polished dashboard make it one of the most beginner-friendly ways to start residential proxy work, and its Site Unblocker pricing ($0.95 per 1,000 requests) is competitive for unlocker-style workloads. For low, stable monthly volumes, the subscription model is predictable and easy to budget.

Where Thordata wins: unit economics at scale and geo precision. The published slider from $2.00/GB down to $0.65/GB with no sales negotiation, ASN-level targeting at no extra cost, and a 100M+ IP pool make it the stronger fit as volume grows past roughly 50–100 GB per month. The gap at 500 GB/month is material: around $500 or less on Thordata’s slider versus roughly $1,375+ on Decodo’s published residential tiers.

At small volumes, both are affordable and the choice is about workflow preference. The divergence is a scaling story, and it’s the kind you want to notice at 50 GB, not at 500.

FAQ

IP pool size doesn’t matter at all? It matters, just less than people assume and differently than marketed. Pool size is a proxy (forgive the pun) for diversity and subnet health. A huge pool concentrated in flagged ASNs underperforms a smaller, well-distributed one — which is why the audit checks ASN diversity directly instead of stopping at the address count.

How do the two compare for scraping-API workloads, not raw proxies? Decodo’s Scraping API runs on monthly subscriptions ($19/$49/$99) with a free starter plan — friendly for low volumes. Thordata’s Scraper API bills per delivered result (~$1.00 per 1,000 at entry, ~$0.50 at scale), which becomes the better deal as volume grows past roughly 100K results monthly. Same crossover logic as the proxies, one layer up the stack.

What about SERP workloads on residential IPs? Don’t run SERP collection on raw residential proxies at all — it’s the worst cost fit, because SERP HTML is heavy and block-prone. Per-response pricing (Thordata’s SERP API from ~$0.70 per 1,000 structured responses) or the managed SERP monitoring solution is the correct architecture regardless of which proxy provider you use for everything else.

Can I audit both providers before choosing? Yes, and you should. Run the 30-line audit above against both trials on your real target domains, at your real expected volume, for a week. Compare delivered results per 100 attempts and cost per delivered result. That comparison takes less effort than arguing about pool sizes on the internet — and it’s the only one that reflects your traffic.

Does either provider require a contract? Neither locks you into a long-term commitment at self-serve tiers — Decodo bills monthly subscriptions, Thordata bills by volume with a 1 GB minimum. Enterprise agreements exist at both when you need SLAs and DPAs.

The Audit Is the Answer

Pool size marketing will keep escalating; the numbers that govern your outcomes won’t change. Diversity, geo accuracy, session control, protocols, sourcing — five checks, one week, your own traffic. Run them against Thordata’s trial (residential trial plus free scraper/SERP API credits), run them against Decodo’s 3-day trial, and if SERP collection is in your mix, evaluate it as its own workload through the SERP monitoring pipeline rather than through raw proxy economics. The provider that wins your audit deserves your traffic — and with published pricing from $0.65/GB at volume, Thordata’s residential network is built to win exactly that kind of test.