Why method matters more than ideas
Anyone can produce a list of business ideas. A list doesn't tell you whether an idea fits you, what it costs to test, what could make it fail or whether the sources behind it are still true. Our method exists to answer those questions honestly, and to make every answer checkable.
How a direction becomes a dossier
Every direction we publish goes through the same five steps before it appears on the site.
- Source. We start from primary public sources: platform and programme descriptions, operator write-ups, official guidance and filings. For each source we record the link, the date we checked it, its language and the geography it describes.
- Mechanic. We describe what actually makes the model work — who pays, what it replaces and how buyers are reached — in our own words. We don't copy closed courses, paid memberships or licensed databases.
- Fit for the target market. We write what changes for buyers in the user's market — price, trust, logistics and rules — and what doesn't transfer at all. A model working somewhere else is context, not proof of demand.
- Counter-evidence. We list local alternatives, reasons the model fails and risks we can name. A dossier without a "why it might fail" section isn't finished.
- Cheapest honest test. We define the smallest test that could change the decision, with its cost, time and a continue condition.
How we decide what's supported
A direction moves from research only to pilot route only when it has a defined process, a quality rubric, measurable limits and an owner on our team who maintains it. Today one route is supported: catalogue data audit as a small service. New routes open one at a time, and only after the previous one shows it can be delivered at quality.
| Status | What it means for you |
|---|---|
| Supported | Full workflow, checks and limits we stand behind |
| Pilot route | Supported for a small group while we measure quality and cost |
| Research only | A dossier to read and learn from — no execution support yet |
How we keep information current
High-impact facts — prices, platform rules, programme status — are rechecked before they're used in a decision. When a link dies or a source changes materially, the affected claim is suspended and the dossier is flagged for review. If a change affects your project, you're told what changed, why it matters and what you can do; general news is never presented as an urgent alert.
How we use AI — and where we don't
AI helps us draft, structure and check material. It runs bounded jobs with structured inputs, and its output is checked before it reaches you: schema validation, deterministic checks for protected fields, valid references for every row or source, and "unknown" instead of invention when information is missing. AI never decides prices, payments, quotas, permissions or whether something is "verified". Text inside a file or a source is treated as data, never as an instruction.
What a label promises
| Label | Our promise | What it doesn't mean |
|---|---|---|
| Fact | A named source and a check date | That it will stay true forever |
| Company claim | Accurately reported from the company itself | That we independently verified it |
| Estimate | The method and assumptions are visible | That the number is precise |
| Hypothesis | Clearly marked as untested | That it's likely or unlikely |
Corrections
If you find an error in a guide or dossier, tell us. We correct it, note the date of the change, and — if the error affected a decision in your project — mark the affected material for review rather than silently changing history.