An SEO Agency’s Honest Math: 50 Clients × Daily Rank Checks × 5 Locations

SERP data is a volume business. An agency tracking 50 clients, 100 keywords each, daily, across 5 locations burns 750,000 SERP requests per month before ad-hoc audits are counted. At typical enterprise rates, that data alone can cost more than the analyst reviewing it. This article runs the real numbers on Thordata and Oxylabs, and explains when each is the right choice.

Search rank data looks cheap until you multiply it by everything an agency actually promises.

The Hidden Multiplication Problem

Ask an SEO account manager how many rank checks they run, and you’ll hear a small number: “About a hundred keywords per client.” Ask their data pipeline, and you’ll hear a much bigger one. Here’s how a mid-size agency’s request volume actually compounds:

That’s 2.25 million SERP requests per month for a 50-client agency — and agencies don’t run one clean request per keyword. Anti-bot systems force retries, CAPTCHAs force re-requests, and location accuracy requires fresh sessions. Real-world inflation of 20–40% on top of the theoretical count is normal.

The uncomfortable conclusion: SERP data cost scales directly with client count, and it scales against your margin.

Why Raw Proxies Make This Worse, Not Better

Some agencies try to escape API pricing by scraping Google directly through residential proxies. For a while, it looks cheaper. Then three costs surface:

1. Maintenance cost. Google’s HTML changes constantly. A parser written in January is broken by June, and someone has to fix it — usually your most expensive employee.
2. Accuracy cost. Without precise geo-targeting and device emulation, ranks drift by position. When a client disputes a report, “our scraper was probably in the wrong datacenter” is not a defensible answer.
3. Reliability cost. CAPTCHAs appear precisely on high-value, high-frequency keywords. The data you most need is the data most likely to be blocked.

This is why the SERP API category exists, and why the right question isn’t “proxy or API” but “which API, at what unit economics.”

The Real Numbers: Thordata vs. Oxylabs

Both Thordata and Oxylabs are credible, established SERP data providers. The difference shows up in pricing structure and what each organization is optimized for. Here are the published numbers:

Running our 50-client agency through both models tells the story. At 2.25 million SERP requests per month, Thordata’s SERP API at its volume rate lands the raw data cost around $1,575–$2,700/month depending on where you sit on the volume curve. An equivalent residential-traffic approach through Oxylabs’ published tiers — the Corporate bucket of 1 TB at $2,500/month — assumes you’re also using that traffic for other things, because raw SERP HTML is heavy. Per-GB pricing punishes SERP work specifically: the same keyword check that costs one API response costs a full HTML page download on a proxy.

To be clear about where Oxylabs wins: if your agency is part of a larger organization that needs a signed enterprise agreement, a formal SLA, named support contacts, and SOC 2 paperwork, Oxylabs’ enterprise motion is built for exactly that, and its network (175M+ IPs) is one of the largest in the industry. That’s a legitimate reason to pay enterprise rates. Oxylabs is a strong product that many large teams rely on.

Where Thordata wins for agencies: transparent self-serve pricing, a per-response model that matches how agencies actually consume SERP data, and a published volume slider that starts at 1 GB with no sales conversation required. For a 15–100 person agency with technical staff that doesn’t need an enterprise agreement, the unit economics are hard to argue with.

A Rank-Tracking Skeleton You Can Ship This Week

Thordata’s SERP API returns structured results — organic listings, featured snippets, ads, and related questions — without you parsing any HTML:

from thordata import Thordata

client = Thordata(api_key="YOUR_API_KEY")

def check_rank(keyword: str, domain: str, location: str, device: str = "desktop"):
    result = client.serp(
        query=keyword,
        search_engine="google",
        location=location,        # country, state, or city level
        device=device,
        output_format="json",
    )
    for item in result.organic:
        if domain in item.get("url", ""):
            return {"keyword": keyword, "location": location,
                    "device": device, "position": item["position"]}
    return {"keyword": keyword, "location": location,
            "device": device, "position": None}

# One client, one keyword, five locations
for loc in ["United States", "New York, NY", "Chicago, IL",
            "Austin, TX", "Seattle, WA"]:
    print(check_rank("crm for law firms", "example-crm.com", loc))

Because the API handles CAPTCHA solving and geo-targeting internally, the response you get is the response you pay for. Agencies that want the collection layer fully managed — scheduled crawls, historical snapshots, alerting — can go one step further with the SERP monitoring solution, which packages continuous SERP data crawling into a monitoring workflow rather than just an API.

Pricing Tactics That Agencies Actually Use

Beyond picking the provider, how you structure collection matters as much as the unit price:

• Tier your keyword sets. Head terms daily, mid-tail weekly, long-tail monthly. Most agencies discover that 20% of keywords drive 80% of client conversations — and cut their data bill accordingly.
• Share the geo matrix. Five locations per keyword is a choice, not a law. Clients in one metro don’t need four other cities checked daily.
• Cache aggressively. Rank data doesn’t change hourly for most keywords. A 24-hour-old rank position is fine for reporting; only alerts need freshness.
• Separate reporting from alerting. Alerting on a handful of money keywords can be real-time; the monthly report can be built from cached data at a fraction of the cost.

These tactics cut the 2.25 million requests down by 40–60% at most agencies — and when the remaining volume is billed per structured response through a cost-transparent SERP monitoring service, the data line item stops being the scariest number in your P&L.

The Bottom Line for Agencies

SERP data cost is a structural expense that grows with every new client. The agencies that scale profitably treat it like infrastructure: they know their cost per client per month to the cent, they use per-response pricing instead of per-GB pricing for HTML-heavy work, and they reserve premium enterprise contracts for when they actually need one.

Start by measuring: take your current keyword count × locations × devices × refresh frequency, and price it against both models above. If your volume looks like the agency in this article, running that calculation with Thordata’s SERP monitoring pricing — with the free trial credits first — will take an afternoon and might fund next quarter’s hiring plan.