Every idea LumenInsight surfaces gets the same treatment: pulled from a discovery source, reduced to the problem it actually solves, then cross-examined against four independent signal families before a single number comes out the other end. That number is the Opportunity Score, and it’s built from three axes.
The three axes
- Friction (30%) — how easy it is to build on or replicate this. Higher means easier.
- Viability (30%) — the technical soundness and sustainability of the project behind the idea.
- Market (40%) — evidence of real demand, drawn from the Evidence Chain.
Market carries the most weight on purpose. Friction and viability tell you whether you can build something — Market tells you whether anyone would care if you did. Demand is also the hardest of the three to fake, which is why it tips the scale.
Why four signals instead of one
One source is an anecdote. Four in agreement is evidence. Every problem statement is checked against what’s being built (GitHub, Show HN, Product Hunt, Betalist), discussed (Hacker News, Reddit), used (npm/PyPI downloads) and searched (autocomplete data). When the chain agrees, the score goes up — when it doesn’t, the report shows that too, rather than smoothing it over.
What we won’t do
We never invent market sizes, and we never claim demand the signals can’t actually show. When the data is thin — a very new repo, a source that failed to respond that day — the score reflects that instead of pretending otherwise.
The full breakdown, including known limitations like English-language bias in our sources and noisy signals for very recent projects, lives on the methodology page. If you disagree with a score, that page — and the raw signals behind every idea — is where to go check our work.