Skip to content
Free scan

Methodology · scoring v1.0.0

How the ZapWeb health score works

A transparent, versioned method. Deterministic checks collect evidence; the score summarizes it. Nothing is estimated that we didn’t measure.

Check statuses

Every check returns exactly one status: pass, warning, fail, not_applicable or unavailable. A page we couldn’t reach is never marked healthy.

Severity weights

Within a category, each check counts by severity: critical 10, high 6, medium 3, low 1, informational 0. A pass earns full credit, a warning half, a fail none. A failed critical check caps its category at 40.

Category weights

The overall score is the weighted average of the category scores shown here. Categories without enough measurable checks (under 40%) are excluded and the remaining weights renormalize, so missing data never drags a score down or props it up.

Coverage and confidence

Coverage is the share of scored checks we could actually measure. Confidence is high at 85% coverage or more, medium from 60%, and low below that. Both appear on every report.

What a public scan can’t see

  • Keyword rankings and Search Console data (require your authorization).
  • Every DKIM selector and your true sending-IP reputation (require message headers or a connected mail provider).
  • Content rendered only by JavaScript — flagged, because many AI assistants and social previews can’t see it either.
  • Real-user Core Web Vitals when Google has too little Chrome traffic for your site.

Heuristics are labeled

Some conversion checks (calls to action, social proof, FAQ content) are hypotheses rather than defects. They’re labeled heuristic and should be validated with real conversion data or A/B tests.

History stays honest

Each report is an immutable snapshot recording the engine and scoring versions. If the methodology changes between two scans, comparisons say so, so a method change is never mistaken for an improvement.

No guarantees

A high score is not a guarantee of rankings, traffic, sales or inbox placement. Google and mailbox providers control those outcomes; ZapWeb measures them.

Our crawler

ZapWeb identifies itself as ZapWebBot/1.0, fetches a bounded number of pages (12 per QuickScan), uses short timeouts, never submits forms, and never probes credentials or ports.