Comparison Methodology

ScrapingPort separates provider-published pricing from calculated comparison values. Public prices come from the latest verified observation in the pricing database and include a source, evidence reference, and verification date.

Default categories

A workload entered in Requests shows scraping APIs by default. A workload entered in GB shows proxy providers by default. These products are kept separate unless the user explicitly enables the combined directional view.

Purchasable monthly cost

Subscriptions use their full minimum monthly charge and are included only when their published allowance covers the requested usage. Annual plans show the monthly equivalent while retaining the annual commitment. Pay-as-you-go packages round up to enough whole packages to cover usage. Tiered plans use the verified tier that contains the workload.

Raw and effective CPM

Raw CPM is estimated infrastructure cost per 1,000 attempted requests. Effective CPM is estimated cost per 1,000 successful requests. A Requests workload represents the desired successful monthly requests. Request and credit products apply provider-published allowances, target rules, and credit multipliers. For proxies, the engine estimates the additional attempts needed at the selected success rate, then converts those attempts to bandwidth using the selected average response size.

Formula: proxy raw CPM = price per GB x response size in GB x 1,000. Proxy effective CPM = raw CPM / expected success rate. Purchasable package minimums and rounding are applied before the displayed workload CPM is calculated.

Bandwidth convention

ScrapingPort uses decimal bandwidth: 1 GB equals 1,000,000,000 bytes and 1 KB equals 1,000 bytes. The active response-size and success-rate assumptions are versioned centrally, returned by the API, shown beside affected results, and editable in Advanced options.

Combined comparisons

Combined API and proxy results are directional infrastructure-cost analysis, not an apples-to-apples comparison of end-to-end value. Proxy estimates exclude scraping logic development and maintenance, retries and orchestration, browser infrastructure and JavaScript execution, CAPTCHA handling, monitoring and incident response, and ongoing operational overhead.

Targets

Each target is calculated independently. Target-specific provider rules take precedence over generic JavaScript multipliers. Multiple targets are not blended unless workload allocation is explicitly supplied.

Verification and history

Prices are stored as append-only observations. A new price does not overwrite an older observation. Unverified observations are excluded from public results, and current pricing is derived from the latest verified valid observation.

Performance benchmarks

Performance results are published only with their target set, date, geography, sample size, configuration, success criteria, and supporting evidence. A user-configured proxy success rate is a modeling assumption and is never presented as measured provider performance.

Commercial independence

Affiliate and sponsorship relationships do not modify calculations or organic rankings. Sponsored placements are labeled separately.