Source
Where did the information come from?
Research methodology
Evidence before ranking. Structure before opinion.
The objective is not to reproduce vendor marketing more neatly. It is to turn technology markets into inspectable information that can support a better buying decision.
Research chain
Every conclusion should be traceable backwards.
Claim
What exactly is being asserted?
Evidence
How strongly does the source support that assertion?
Canonical data
What can responsibly enter the structured model?
Comparison
How does the supported information affect a real decision?
Four layers
The system is designed to resist confident-looking noise.
Canonical structure
Organizations and products are separate entities. Category, industry, function, use case, capability and technology are independent dimensions.
Source hierarchy
Authoritative and first-party sources are preferred for verifiable facts; independent sources are used where they add evidence a vendor cannot provide about itself.
AI-assisted, not AI-trusted
AI may extract, normalize and classify information. Its output is a candidate claim, not canonical truth.
Commercial independence
Vendor claiming, verification, sponsorship and organic research are separate states. Payment does not silently change a conclusion.
Missing is not false.
If a material fact cannot be supported, the correct state is unknown or unverified. Absence of evidence should not be converted into a negative claim merely to make a comparison look complete.
The model must be falsifiable.
Technology data changes quickly. Better evidence should be able to replace weaker evidence. Corrections and challenges are welcome at belobrovtech@protonmail.com.
See it applied