Most idea tools hand you eight numbers and no way to rank them. IdeaWave gives you one: the IdeaWave Score (0-100), paired with an Evidence Score that caps how high a thinly-sourced idea can ever go. A confident-looking idea built on thin data physically cannot post a high score.
Each sub-score (1-10) is anchored to real signals (search volume and growth, community pain, verified revenue from comparable businesses), then blended with these weights:
| Problem / pain severity | 25% |
| Why now (timing) | 20% |
| Opportunity | 15% |
| Go-to-market | 15% |
| Feasibility | 10% |
| Revenue potential | 10% |
| Value equation | 5% |
Pain leads the formula because it's the single best predictor of a sellable product. Why-now is weighted heavily because it's both our most data-anchored signal and the one that decays, so the feed stays fresh.
Before weighting, the whole composite is multiplied by an evidence penalty (0.70-1.00) computed from citation coverage, signal completeness, source quality, and whether payment-verified revenue comparables exist. That's why you'll see scores like IdeaWave 82 · Evidence 91: one number you can rank by, and a second that tells you how much to trust it.
Revenue data enters the system through five source channels grouped into four trust tiers, and each row carries its provenance everywhere it appears:
Payment-verified
Pulled from platforms that read revenue directly from Stripe, Lemon Squeezy, Polar, or RevenueCat. The strongest evidence tier: it counts fully toward the Evidence Score and is the only tier that can push revenue potential to HIGH on its own.
Self-reported
Founder-posted figures collected from public build-in-public posts. Useful signal, lower weight. When a monthly figure is just an annual number divided by twelve, we label it "Est. (ARR/12)" instead of pretending it was measured.
Modeled (App Store)
Apple publishes no revenue numbers, so every app-revenue figure anyone shows you is an estimate. Ours are computed, not licensed: the Top Grossing chart is Apple's own revenue ordering, and we fit a rank-to-dollars curve through it, cross-checked against ratings momentum, paid-chart position, and payment-verified anchors. Modeled figures publish as a single point estimate with the full modeled range one hover away and itemized in the estimator breakdown; they never wear the verified seal, never rank on the startup leaderboard, and an app without enough signal shows no number at all.
Synthetic
Development-only placeholder data. It earns zero evidence credit, can never be labeled as proof, and the publish gate rejects any idea built on it outright.
In plain English: we capture Apple's US charts every morning, fit a curve through six fixed rank checkpoints, require it to always go down, and keep it smooth. The curve's overall level is pinned by published App-Store-wide spending totals, and category charts are placed against the overall chart using apps that appear on both. Independent reads then sharpen that number: how fast an app collects ratings (priced per rating for its monetization type), where a paid app sits on the Top Paid chart times its price, the divergence between an app's free and grossing positions, and — strongest of all — a payment-verified revenue figure when one of our verified startups turns out to be the app's maker. Every signal is weighted by its own uncertainty and fused into one published point estimate. When signals disagree, the modeled range around the point widens; when signal is too thin, we publish nothing.
The starting level of the curve is seeded from public third-party figures (which are themselves estimates), and we say so: early on, our numbers partly measure whether we copied the industry's level correctly. The model recalibrates weekly against a permanent hold-out set of verified figures it is never allowed to fit, and a new calibration only ships if its hold-out accuracy did not get worse. Once a month we blind-test against third-party estimate services with our numbers frozen in advance, and publish the agreement statistics here — including the bands where we look bad.
One honest caveat: Apple's grossing rank is revenue-ordered but not purely so (the chart has velocity weighting and anti-gaming adjustments baked in). That is the reason the curve must be fitted and error-checked rather than assumed — and the reason every point estimate on this site carries its full modeled range in the breakdown, never just a naked number.
Every community signal links to a real, dated source. Every leaderboard row links to the actual product. Trust is the product.
Pro · All of this scoring exists to point at one thing worth building. Watch one get built.