Evidence & sources
IdeaSweetSpot's coaching prompts and journey stages are grounded in a curated knowledge base of 291 principles about validating, scoping, building, and launching small products. Every principle was harvested from public research and practitioner sources, checked against its original source, graded for evidence strength, and de-duplicated into one canonical library. This page explains what those labels mean — and the rules the coach never breaks.
The short version: the coach cites what it leans on, says so when something isn't well established, and never invents facts, statistics, or citations. Your scores are your own honest read — we don't scrape markets or generate "data" about your idea.
The trust ladder
Every principle carries a plain-language trust label based on the strength of the evidence behind it. You'll see these labels wherever the coach cites a principle:
| Label | What it means | Count |
|---|---|---|
| Strongest evidence | Backed by rigorous research — meta-analyses, replicated studies, or converging findings across fields. | 79 |
| Strong evidence | Well-supported by research or consistent, credible practitioner data. | 158 |
| Practitioner rule of thumb | Widely used by experienced builders and coaches; useful, but not formally proven. | 51 |
| Emerging | Promising but early — treat as a hypothesis to test, not a rule. | 3 |
What the library covers
The 291 principles span nine domains, so the coaching can connect the business read with the human one: entrepreneurship and validation, software engineering, economics and pricing, product and UX, design, growth, behavioral science, organizational psychology, and executive-coaching practice (the ask-first, GROW-style stance the coach takes with you).
How a principle earns its place
- Harvest: candidate principles are gathered from public research and reputable practitioner sources.
- Adversarial source-check: each claim is traced back to its origin. Claims that turn out to be misquoted, overstated, or unsourced folklore are cut or downgraded — several popular startup "statistics" didn't survive this step.
- Grading: what survives is tiered by evidence strength, which becomes the plain-language trust label you see.
- Deduplication: overlapping findings are merged into one canonical card, so the coach doesn't double-count weak evidence as two sources.
What the coach refuses to do
- No invented facts, statistics, market sizes, or citations — the grounding instructions travel with every prompt, including the ones you paste into your own AI.
- No scraped or generated "market reports." The brief gives you a market-signal checklist to run yourself, so the evidence is yours and stays on your device.
- No score inflation. The verdict comes from your own four-circle read, and the weakest circle is always surfaced — the candor is the point.
- No verdicts handed down. The coach asks before it tells, and the build / don't-build call is always yours.
The boundary
IdeaSweetSpot is a reflective prioritization tool. It is not financial, legal, career, tax, or mental-health advice, and nothing in it guarantees income, demand, or success. Everything here is the author's opinion, for general information — when in doubt, consult a qualified professional.
Questions
Want to know how a specific principle is sourced, or spot something that looks wrong? Email mthibideau@apps4that.com — corrections are welcome.