Research methodology

Submit an exact product link and destination, an established foreign business, or a discovery brief. The researcher checks identity, target-market availability, competitors, supplier terms, demand, costs and rules. The current initial search stage permits up to six web-tool calls; a call can cover several pages. It chooses one candidate assessment rather than promising exhaustive market coverage.

Sources come from recorded retrieval and bounded public page/PDF capture. Static product metadata is advertised publisher information. The review checks atomic statements against short literal source quotations. Failed capture, search-only evidence, contradictory text and omitted review items remain visible. A domain count is not a count of independent companies, and an automated source check is not a human investigation.

Reject means the evaluated version lacks a defensible case under the stated scope; it can be a useful result. Needs more validation means material evidence remains unresolved. A possible niche is a research hypothesis, not proven demand. Read the scope, open checks, evidence age and any conflicts. Neither absent search results nor a foreign brand’s delivery restriction proves a missing local category.

New reports record whether useful checks were covered, further passes produced no new relevant results, access was blocked, a search budget was reached or critical questions remain open. Older reports may not contain a saved stopping reason. Follow-ups can address one missing check; they do not automatically rerun or overwrite the original report.

Your research stays in your account. Public examples must be explicitly selected and checked for confidential information before publication. Fictional walkthroughs are labelled fictional; curated source-backed examples explain their evidence limits and dates. We do not claim a public example establishes an unoccupied or profitable market.