Benchmarks
Ecommerce scraping benchmarks
Benchmark pages are only useful when they are repeatable and conservative. This comparison is designed around public targets, repeated runs, and practical ecommerce monitoring metrics: successful extraction, speed, block rate, and effective cost per successful page.
Latest benchmark run
Generated 3 June 2026 at 14:29. Each provider was tested 5 times per target.
| Provider | Success rate | Median duration | Effective cost / successful page |
|---|---|---|---|
| ScrAPI | 100.0% | 3951 ms | $0.00095 |
| ScraperAPI | 50.0% | 20282 ms | $0.00490 |
| ScrapingBee | 50.0% | 4692 ms | $0.00098 |
| ScrapingAnt | 50.0% | 6628 ms | $0.00190 |
| Zyte | 75.0% | 6367 ms | $0.00300 |
| Firecrawl | 75.0% | 3450 ms | $0.00320 |
What we measure
A run counts as successful only when the provider returns a usable page and the expected text is present. HTTP 200 alone is not enough, because captcha pages and block pages often return successful status codes.
What we avoid
The benchmark avoids customer URLs, private targets, login-protected pages, and one-off claims that cannot be reproduced. The goal is a fair comparison, not a takedown of other providers.
How to read it
Success rate and block rate matter more than headline request pricing. For ecommerce monitoring, the useful unit is a successful price or stock observation delivered at a predictable cost.