Pay-per-row pricing explains the model: you pay for rows returned, not requests made. What it doesn't tell you is that a "row" costs wildly different amounts depending on which endpoint you're calling. This post is the numbers — what a real run actually costs, across enough verticals to see why the range is so wide.
Why does the same "one row" cost 400x more in one vertical than another?
The cheapest endpoint in the catalog is Indeed Jobs at $0.00005 per job. The most expensive per-unit is LinkedIn Email at $0.02 per email — 400 times more for what sounds like a comparably-sized "one row."
The gap isn't arbitrary. A job listing is already a structured public page — the site did the hard work of laying it out cleanly, and returning one row is close to a straight extraction. A work email address isn't sitting on any public page at all: turning a LinkedIn profile URL into a verified email means running an actual lookup, not just reading a page that already has the answer on it. Price tracks effort, not row count.
What does a real 500-row LinkedIn email run actually cost?
500 profile URLs through LinkedIn Email at $0.02/email: $10. Feed the same 500 URLs through LinkedIn Profiles instead, at $0.004/profile: $2. Same input size, 5x apart, because one endpoint returns a public profile and the other verifies a private contact detail.
This is the number worth knowing before you build a pipeline: enrichment steps that resolve something not already public cost meaningfully more per row than lookups that return what a page already shows. Budget for that difference deliberately instead of discovering it in your first invoice.
What does a real 1,000-row review pull cost?
Reviews are near the cheap end across the board, but even within "cheap" there's a real spread. 1,000 rows through Amazon Reviews at $0.0005/review: $0.50. The same 1,000 rows through Google Play Reviews at $0.0001/review: $0.10 — the single cheapest endpoint in the entire catalog. Both are plain-text extraction from an already-public review thread; the gap here is closer to noise than a meaningful signal, which is itself useful to know: don't read too much into small differences at this end of the price range.
What does a cheap, high-volume pull look like end to end?
Take Indeed Jobs at $0.00005/job. A 5,000-job pull — enough to cover a metro area's postings for a mid-sized role search — costs $0.25. Compare that to a single TikTok Shop pull of 500 products at $0.015/product: $7.50, for a tenth of the row count. Volume and price-per-row move independently of each other; a "big" pull in row count isn't necessarily an expensive one, and a small one isn't automatically cheap.
| Endpoint | Price | Example run | Cost |
|---|---|---|---|
| LinkedIn Email | $0.02 / email | 500 emails | $10.00 |
| TikTok Shop | $0.015 / product | 500 products | $7.50 |
| LinkedIn Profiles | $0.004 / profile | 500 profiles | $2.00 |
| Amazon Reviews | $0.0005 / review | 1,000 reviews | $0.50 |
| Google Play Reviews | $0.0001 / review | 1,000 reviews | $0.10 |
| Indeed Jobs | $0.00005 / job | 5,000 jobs | $0.25 |
How do you estimate cost before you run a job?
Multiply the per-row price by how many rows you actually expect back — not
how many you're willing to accept as a ceiling. A search-style endpoint with
a maxResults cap bills only what it returns, so setting that cap
generously doesn't cost you anything until rows actually come back. The
number to estimate honestly is your realistic result count, not your
worst-case limit.
For a recurring job, multiply that per-run cost by how often it runs. $0.25 for a 5,000-row Indeed pull looks trivial once; run it nightly for a month and it's $7.50 — still cheap, but worth knowing before you set up the schedule, not after thirty invoices.
A multi-step pipeline chains more than one endpoint together, and that's where the per-row differences above actually compound. Take an ABM run into 50 target companies: LinkedIn Companies at $0.004/company to confirm each account (50 × $0.004 = $0.20), then LinkedIn Company Employees at $0.004/profile to pull roughly 20 people per company (1,000 profiles × $0.004 = $4.00), then LinkedIn Email at $0.02/email to resolve contact details for the 200 people who actually match your target roles (200 × $0.02 = $4.00). Total: $8.20 for a fully worked account list with verified emails — and more than half of that cost sits in the email step alone, even though it's the smallest row count in the chain. That's the pattern worth internalizing: the expensive line in a pipeline is rarely the one with the most rows, it's the one doing the most work per row.
FAQ
Profiles returns what's already public on a profile page. Email resolves a work email address that isn't published anywhere — that's a verification step, not a read, and the price reflects the extra work.
No. You're billed for rows actually returned, not the cap you set. Setting a high cap costs nothing extra unless that many rows genuinely exist to return.
No — price reflects how much work the data required, not how useful it is. Pick the endpoint that returns the fields your job actually needs; comparing prices across endpoints that return different data isn't a fair comparison.
