Automated scan (raw evidence)
Tools measure speed (LCP/CLS/INP), links, HTTPS, sitemap, robots.txt, images, headers and HTTP status. Raw results are stored separately from any interpretation.
Metodología / Methodology
How the report is built
Every report follows the same verifiable structure. This page explains the method; the full sample report shows it in practice.
Human review (findings)
A person walks the site like a real customer: forms, purchase/booking, trust, accessibility, mobile, local competitors. Every observation becomes a finding with attached evidence.
Prioritisation by impact
Each finding is scored on business impact (customers/sales/risk) and fix effort. The resulting matrix sets the action-plan order: first what hurts and is cheap to fix.
A plan with “done means…”
No task enters a sprint without a clear, measurable acceptance check. That way nobody argues later about whether something is finished.
§ Severity
Three levels, defined by impact on customers and business — not by aesthetic preferences:
Critical — you lose customers or money right now Warning — it keeps costing you sales or trust Improvement — it adds up, but it isn't bleeding
§ The five-piece rule
An incomplete finding never gets published. Every finding carries: what happened · why it matters · evidence · recommended action · “done” means…. Raw evidence stays separate from our interpretation so you can check it yourself.
§ What the full report includes
- Executive summary — global score + breakdown + 5 priority actions
- Detailed findings — one by one, with all five pieces
- Impact × effort matrix — the recommended fix order
- Raw evidence — measurements, screenshots and logs, uninterpreted
- Action plan — suggested sprints with “done means…”