Methodology
How opportunities are assessed
One rubric, ten qualitative signals, mandatory human approval. Here is exactly how an assessment gets published — and what we refuse to do.
The rubric
Every opportunity is rated on the same ten signals, and every rating is expressed on one scale: Favorable, Mixed, or Unfavorable. Seven are qualitative factors: marketing-channel feasibility, gross-margin potential, time to first revenue, stickiness and retention, build and technical complexity, founder interaction and service posture, and scalability. Three are market-dimension ratings: Customer Economics (a Favorable/Mixed/Unfavorable grade with a free headline figure plus a paid full breakdown), Competitive Intensity, and Disruption Risk.
Competitive Intensity and Disruption Risk are measured levels, recorded as Low, Medium, or High. On public pages they are translated onto the same grade scale — Low reads Favorable, Medium reads Mixed, High reads Unfavorable — with the measured level shown alongside the grade so no information is lost.
We deliberately publish no numeric scores on free pages. Decimal-point scores imply a precision that qualitative judgment does not have, and they invite ranking opportunities in the abstract rather than against your own constraints.
The editorial process
- Research. Every factual claim is stored with stable source records and the exact draft identity it supports. Editors separately record source health and confirm that the source actually supports the claim; a reachable URL alone is not evidence approval.
- Draft. A draft revision is written — AI may assist with drafting, but AI output is never published directly.
- Review. A versioned gate checks evidence coverage, specificity, repetition, qualified claims, opportunity boundaries, catalog collisions, and discoverability. SEO may honestly be marked not viable; keyword demand never overrides the durable opportunity.
- Approve & publish. Authenticated staff see the findings and retain both decisions. Approval and publication run the same complete validator against the same policy and content/evidence hash. Any edit invalidates the prior assessment and approval. Published revisions are versioned and immutable — paid snapshots always reference the exact revision you purchased.
What this is not
These assessments are qualitative editorial judgments, not investment, legal, or financial advice. They cannot account for your capital, skills, network, or risk tolerance — the comparison exists precisely so you apply your own constraints to consistent information.
Ready to see it in practice? Browse the directory or read the free guides.