Laboratory or field data?

Both where available: own laboratory measurement in the real browser for each tested page, supplemented by public field data of the domain if Google provides it.

My hoster says the site is fast — why check?

Server response time is only a fraction. LCP, CLS and responsiveness are created in the browser — through images, fonts, scripts and layout. That's where we measure.

What is the current outcome?

Each seal shall bear the date of verification and its validity; the audit is a documented snapshot. Tests may be repeated for ongoing inspection.

This translation has not been editorially reviewed yet. The German version is authoritative. German version →

Loading performance: what exactly is checked

Core Web Vitals (cwv-v1). Every subtopic leads to its checkpoints with weight and explanation; the list comes from the audit engine itself.

Charge time
LCP, FCP, TTFB and Speed Index: when the content is there.
4 articles
Response, stability and measurement
INP, Total Blocking Time, CLS and the difference between field and laboratory.
4 articles

Bildnachweise

Charge time: Suohros, CC0
Response, stability and measurement: William Stitt, CC0

What is measured

This audit area rates from findings and measured values, not from a countable set of rules. That is why this page lists which signals are collected.

Charging behaviour (LCP)
When the largest visible content appears — the moment when the page is “there” for visitors.
First Content (FCP)
When the first pixel content appears. The value separates “Server is slow” from “Representation is slow”.
Response time (INP)
How long it takes the page to visibly respond to an input — across all interactions of a visit.
Blocking time (TBT)
How long the main thread is so busy during loading that entries remain.
Layout stability (CLS)
Whether content leapfrogs during loading — the cause of accidental clicks on the wrong item.
Server Response (TTFB)
How long it takes for the server to reach the first byte. Everything else can only begin after that.
Speed index (SI)
How quickly the visible area fills up — a gradient instead of a single point in time.
Field data of real visitors
Values from the 28-day window of actual calls, if there is enough data for the address.
Laboratory data under fixed conditions
A synthetic barrel with a throttled joint and defined apparatus — comparable but not representative.
Mobile and Desktop Separated
Both strategies are measured independently; A good desktop value says nothing about the phone.