UV Skinz · evidence-led Lighthouse history

Audit evidence, visible over time.

Every supplied Lighthouse record remains visible. Metric charts use only captures with an explicitly recorded value; source, status, version, and time precision remain visible so unlike evidence is never presented as a single benchmark.

Comparison guardrail: live-storefront, preview, uploaded PageSpeed data, narrative medians, Lighthouse versions 13.1.0–13.4.1, and failed HTTP 410 captures are not interchangeable. Filter to a comparable source and route/device before interpreting movement.

Explore score-bearing observations

Dates use their actual spacing when an exact timestamp exists. September 2 narrative medians have an explicitly date-only time marker; error records remain in the register and coverage view, not the metric chart.

Performance score by audit date

Coverage and collection status

This is an audit-record view. It includes failed and unavailable captures without replacing unavailable metrics with zero.

September 2 matrix context

Narrative-level five-run medians documented in the supplied matrix README. These rows have no attached raw JSON in the 54-row master and are displayed as summary evidence, not individual captures.

Route / device	Median score	Median LCP	LCP element across runs	Status

Unattributed May 17 / May 18 comparison

The newly supplied 72-row audit comparison has no URL, year, device, Lighthouse version, or run identity. It is intentionally kept out of the dated run chart and shown here only as an unattributed historical artifact.

Do not use this table to claim a performance regression, improvement, route result, or day-to-day trend until its raw exports or an auditable run manifest are supplied.

Audit	May 17 displayed value	May 18 displayed value	Record status

Complete 54-record observation register

The master report is embedded exactly once. Each record preserves its supplied source, evidence state, and metric availability.

Time	Source	Route / device	Status	Score	LCP	Runtime	Evidence