AI News

How the Kingy Launch Score works

The Kingy Launch Score rates a product launch from 0 to 10 on how well it is evidenced — not on hype. A score is only shown once a launch clears a sourcing floor; until then it reads “Needs review.” No launch is scored by default, and no launch gets a 10 for showing up.

What the score measures

Every score is the weighted average of five judgments. We never assign a final number by hand or by feel — each part is judged against the launch’s own cited sources, and the total is computed from those parts.

Part Weight What it asks
Source & verification quality 25% Are there official, dated sources we can check independently?
Product evidence 25% Is there a real demo, repo, API, or docs — or only an announcement?
Significance & novelty 20% How much does this move its category, compared with what already exists?
Traction signals 15% Is adoption, usage, stars, or funding verifiable and cited?
Offer clarity 15% Are pricing, access, and terms actually documented?

The first two parts are a floor. If a launch has no checkable sources or no real product surface, it cannot receive a number — it stays “Needs review” until an editor resolves it.

What the bands mean

Score Band What it says
9.5–10.0 Exceptional Category-defining and fully sourced. Capped at the top ~2% each quarter and signed off by a named editor.
8.5–9.4 Strong High quality, well evidenced.
7.0–8.4 Solid Real, useful, and sourced.
5.0–6.9 Mixed Partial evidence or meaningful gaps.
Below 5.0 Weak Thin, self-reported, or announcement-only.

The rules we hold ourselves to

  • Sourced or unscored. If a fact isn’t in a source we can link, it doesn’t affect the score and it doesn’t appear in the write-up. We would rather show “Needs review” than guess.
  • Any 9.0 or above is signed off by a person. Curtis or Gilbert reviews it, and the page shows who reviewed it and when.
  • Every score keeps its history. When a score changes, we record the old value, the new value, the parts behind it, the date, and why. Nothing is overwritten silently.
  • The write-up is original. The short “Kingy take” on each launch is written for that launch, cites its source inline, and never repeats the same boilerplate across records.

Why scores changed in July 2026

We rebuilt this system because our earlier scores had drifted: too many launches clustered on a handful of identical numbers, and some carried a placeholder 10 that was never a real judgment. In July 2026 we reset every launch to “Needs review” and began re-scoring them under this methodology (v2), rewriting the affected notes as we go. A full before-and-after accountability post is on the way.

Methodology v2.0 · last updated 2026-07-09 · questions: curtis@kingy.ai