How the score works
The full formula, the weights, why each dimension matters — and the honest limits of what a 0–100 score can tell you.
The formula
Every question has four possible answers, each worth points:
| Answer | Meaning | Points |
|---|---|---|
| Ready | Fully in place and verified | 1.0 |
| Partial | Started or in progress, not launch-ready | 0.5 |
| Missing | Not done — becomes a blocker | 0.0 |
| N/A | Doesn't apply to this launch | excluded |
Step 1 — dimension score. For each dimension, add up the points of answered questions (N/A questions are excluded entirely) and divide by the number of answered questions, then multiply by 100 and round to the nearest whole number:
dimension score = round(100 × points ÷ answered questions)
Step 2 — overall score. Multiply each dimension score by its weight, add them up, and divide by the sum of the weights:
overall = round( Σ (dimension score × weight) ÷ Σ weights )
If every question in a dimension is marked N/A, that dimension drops out and the remaining weights are re-normalized (divided by their own sum, not by 100). If you somehow mark all 24 questions N/A, the score is 0 — a launch with nothing applicable to check isn't a launch we've measured.
Dimensions and weights
The six weights sum to exactly 100. Technical and Launch mechanics carry the most weight because they're where launches actually die: an unreachable site, or a launch process that was never planned.
| Dimension | Weight | Why it matters |
|---|---|---|
| Technical | 20% | Hosting, DNS/TLS, backups, rollback. If the site is down or unrecoverable, nothing else matters. |
| Launch mechanics | 20% | The launch itself is a project: a written staging→production checklist, flagged unfinished work, comms, tested critical flows. |
| Content | 15% | Finished copy, legal pages, error states, working forms. This is what visitors actually see. |
| Discovery | 15% | SEO basics, sitemap, analytics, social previews. A launch nobody can find is a tree falling in a forest. |
| Trust | 15% | Contact route, privacy posture, support path, business identity. Trust is the conversion rate's quiet multiplier. |
| Post-launch | 15% | Monitoring that alerts a human, error tracking, on-call, feedback loop. Launch day is the start, not the finish. |
Worked example: N/A handling
Say your Technical answers are: hosting ready, DNS/SSL ready, backups partial, rollback plan N/A (you're launching a static brochure site with nothing to roll back).
- Points: 1.0 + 1.0 + 0.5 = 2.5
- Answered questions: 3 (the N/A is excluded from the denominator)
- Technical score: round(100 × 2.5 ÷ 3) = 83
The N/A neither helps nor hurts — it simply doesn't count. Mark N/A honestly: if you're unsure whether something applies, it probably does.
Blocker ranking
Every missing answer becomes a blocker, ranked by launch risk:
- Critical — can sink the launch outright (no rollback plan, no monitoring, untested checkout).
- High — likely to cause real damage or lost conversions (no backups, broken contact route, unfinished legal pages).
- Medium — degrades the launch but won't sink it (missing social previews, incomplete analytics).
Within a risk tier, blockers sort by dimension order, then question order. Partial answers are listed separately as "tighten before launch" — they're warnings, not blockers.
Score bands
| Score | Band | What it means |
|---|---|---|
| 90–100 | Ready to launch | No structural blockers found. Run the 48-hour preflight, then ship. |
| 75–89 | Nearly ready | Clear the blockers, re-run the score, confirm every critical item is ready. |
| 60–74 | Gaps to close | Real gaps remain. Work the blocker list top-down before setting a date. |
| 40–59 | Not yet | Too many launch risks open. Treat the blocker list as your pre-launch plan. |
| 0–39 | Hold the launch | Launching now would be flying blind. Start at the top of the list. |
Honest limits — what the score does NOT guarantee
The score is a readiness signal, not a guarantee. A 100 does not mean your launch will succeed — it means you checked the boxes that most often cause launches to fail.
- It's self-reported. "Ready" means you believe it's ready. The score can't verify your backup restore actually worked or your checkout flow truly handles edge cases.
- It's not QA. It doesn't test your code, your load capacity, or your security posture. It's not a substitute for QA, penetration testing, or a security review.
- It's not legal advice. The legal-pages question checks existence, not correctness. Regulations vary by jurisdiction and business — get real counsel where it matters.
- It doesn't measure demand. A perfectly launched product nobody wants still fails. The score measures launch execution, not product-market fit.
- Weights are judgment calls. They're grounded in real launch postmortems, but they're our judgment — reasonable people could weight differently, and no weighting captures every launch type.
Determinism
The same answers always produce the same score — there is no randomness, no server, and no hidden adjustment. The scoring code is a single static file; the formula above is its complete specification. Your answers are stored only in your browser's local storage and never transmitted anywhere.