Why methodology matters

Benchmark results are only useful if the testing process is consistent. This guide provides a practical framework for sampling, measuring, and comparing proxy performance so that results are repeatable and defensible.

What this guide includes

• A consistent sampling approach
• Standard metric definitions
• Tips for avoiding biased results
• A readiness checklist before you run any test

How to use it

Use this methodology before running any benchmark or comparing proxy pools across regions. Capture a baseline, then change one variable at a time to isolate impact.

Recommended Baseline Setup

• **Destinations**: Choose a fixed set of target categories (e.g., ecommerce, search, social)
• **Concurrency**: Start with a low, stable baseline
• **Time window**: Use the same time window for each run
• **Session mode**: Keep sticky vs rotating consistent per run

Interpreting Results

• Compare success rates relative to your baseline, not a single run
• Track latency distribution (p50, p90) instead of only averages
• Document anomalies such as regional outages or transient blocks

Reporting

Summarize results with your configuration, time window, and destination set so that teams can reproduce the test later. Include a short summary of what changed between runs and why, plus any external factors that may have affected outcomes. Document the test environment to make comparisons fair.