Research Methodology
Our methodology is designed to produce inspectable research. Criteria are fixed before conclusions are drawn, evidence is classified consistently, and every report discloses its limitations.
Four-Stage Research Process
01
Signal Capture
Identify evidence asymmetry, market signals, and publicly accessible sources that can be verified.
02
Criteria Design
Define evaluation criteria and weights before assessing any provider, claim, or outcome.
03
Evidence Assessment
Apply criteria consistently, label every claim as Verified Proof, Unverified Assertion, or Proof Gap.
04
Published Output
Publish findings with source notes, limitations, disclosure, and an open correction pathway.
Coverage Gating Tests
We only launch coverage in a category when it passes four tests:
- Evidence asymmetry: claims differ enough from observable reality that independent research adds value.
- Accessible public sources: we can verify material claims without privileged access.
- Bounded, answerable question: the research question can be answered within a defined scope.
- Feasible evaluation framework: we can design criteria that are specific, measurable, and replicable.
Seven Evidence Classes
1. Direct Documentation
Pricing pages, API docs, contracts, white papers, and other first-party materials.
2. Independent Reviews
Third-party audits, customer testimonials with named sources, and verified case studies.
3. Market Signals
Search demand, job postings, practitioner discussion, and conference activity.
4. Regulatory & Legal Records
SEC filings, court records, patent databases, and government enforcement actions.
5. Expert Interviews
On-the-record conversations with practitioners, operators, and academics.
6. Quantitative Data
Datasets, benchmarks, and reproducible analyses from credible sources.
7. Editorial Analysis
Synthesized interpretation, clearly labeled as such and bounded by evidence.
Claim Classification
Every material claim in our reports is labeled using one of three tags:
Evidence Classifier Demo
Every claim in our reports is labeled using this taxonomy.
Eight Standard Criteria Dimensions
For category benchmarks, we evaluate providers across eight dimensions. Weights are set before any provider is named.
- Category fit — does the offering match the category definition?
- Scope clarity — are boundaries and exclusions explicit?
- Proof quality — how strong is the supporting evidence?
- Pricing transparency — are prices public and comparable?
- Implementation model — is the delivery model documented?
- Support & accountability — what support, SLA, and escalation paths exist?
- Trust signals — what independent validation is available?
- Risk factors — what could invalidate the conclusion?
Update Cadence
Live reports are reviewed quarterly or when material new evidence emerges. Reports are date-stamped for publication and last review. Readers may request updates via Submit Evidence.