# Core Panel product-comparison methodology

Version 1.0 — 2026-09-20

## Source hierarchy

1. Reproducible first-party measurements with raw results, environment manifests, runner source, and checksums.
2. Current official vendor pricing and documentation.
3. Core Panel product documentation and source-linked security disclosures.
4. Clearly labelled calculations derived from the sources above.

Unattributed reviews, search snippets, reseller marketing, and unverifiable performance claims are not treated as evidence.

## Pricing

Record the vendor, plan, billing unit, account or server allowance, currency, billing interval, tax treatment, URL, and observation date. Preserve different currencies until the reader supplies a current exchange rate. Compare annual commitments with the same commitment period. Separate license price from infrastructure, extensions, support, payment fees, migration, and labor.

## Capabilities

Compare a named workflow, not a generic checkbox. State prerequisites, product version, platform restrictions, required extensions, and whether the evidence is documentation, a product test, or direct observation. Supported does not mean identical operation or compatibility with every extension.

## Performance

A comparative performance claim requires equivalent hardware, operating systems, workload source, datasets, warm-up, run count, duration, concurrency, software versions, service configuration, failure handling, and statistical summaries. Publish every run and disclose unsuccessful runs. Do not compare one product's measured result with another vendor's marketing number.

## Migration

Report archive composition and size, file count, database and mailbox sizes, source and destination versions, network path, transfer time, conversion time, restore time, validation time, cutover window, downtime, warnings, failures, retries, and rollback outcome. Synthetic fixture measurements must not be presented as customer migration speed.

## Security

Distinguish developer regression tests, automated scanners, independent assessment, and production incident evidence. Publish the tested revision, tool versions, scope, commands, results, and limitations. Passing tests do not establish absence of vulnerabilities.

## Corrections and updates

Every dataset has a stable identifier and date. Corrected artifacts receive a new version or dated directory; do not silently replace a result while retaining its checksum. Material methodology changes update this document's version.