Sbirwatch
A public-data diligence workspace and API that resolves companies, people, awards, spending, and patents across official sources, calculates reproducible program-history indicators, and exposes uncertainty, corrections, and primary records for qualified review.
The research confirms free SBIR, federal spending, and patent interfaces plus public discussion of repeat award recipients. It found no reviewed SBIR watchdog product; a scraper is the closest narrow tool. No structural incumbent copying cost is evidenced.
Sbirwatch must not call a company, principal investigator, award, application, or conversion pattern fraudulent, wasteful, noncommercial, noncompliant, risky, or unqualified from public data or a low Phase I-to-II rate. Agency missions, award types, program rules, time windows, company identity, acquisitions, name changes, research goals, national-security constraints, publication lag, and commercialization paths differ. Absence of a patent, follow-on award, or linked spending record is not evidence of absence.
Every indicator needs a source, query time, pagination and coverage record, entity-resolution method, denominator, cohort, exclusions, confidence, and correction path. Public pages should use neutral language and link primary records. Fraud or enforcement conclusions remain with authorized agencies, inspectors, courts, and qualified investigators.
Venture investors, prime contractors, agency program teams, journalists, researchers, and diligence providers needing traceable SBIR history.
Accessible federal APIs make the product feasible now.
Investors, contractors and program teams have concrete diligence uses.
Several public sources and discussion support the need without a product category.
Three free official interfaces, a clear diligence buyer, hard entity resolution, reproducible indicators, primary-record links, and correction workflows create a practical data product.
The market may be narrow, public data is delayed and incomplete, adverse labeling creates serious risk, conversion is not commercialization, and no structural moat beyond data quality is evidenced.
Discussion
No comments yet — be the first to weigh in.