Method · 4 min read · 2026-10-06

A report template for provider comparisons

What a proxy comparison should record so another team can repeat it: protocol, targets, definitions, samples and raw data.

A comparison is only useful later if someone can tell what was compared and how. This template is the minimum we would want to see in a report, from any vendor or any team, ours included.

Header

  • Dates and times of the runs, with timezone, and the run order.
  • Product and plan tested for each provider, and the session mode used.
  • Client software, version and configuration, including timeouts and retries.
  • Where the client ran, because network location changes timing.

Per target

  • The URL pattern and why it was chosen.
  • The content check that defines success, applied identically to every provider.
  • Requests attempted per provider, and the number that passed.
  • Median and 95th percentile time to first byte.
  • Bytes per success, including failed attempts.
  • Failure mix: 429, 403, soft blocks and timeouts.

Raw data

Publish one row per request, so a reader can recompute anything in the summary and check your arithmetic.

results.csv: one row per request
run_id,provider,target,started_at_utc,status,passed_check,ttfb_ms,bytes,attempt

Test it on your own targets.

Create an account, run the measurement harness against your real workload, and read the numbers before you commit to anything.