This translation has not been editorially reviewed yet. The German version is authoritative. Deutsche Fassung
Performance
What is LCP, and why should it stay under 2.5 seconds?
LCP (Largest Contentful Paint) measures the time from opening a page to the moment its largest visible element is in place on screen — Google rates up to 2.5 seconds as good, above 4 seconds as poor. For visitors, that's the moment the page feels "there". Before that, they see a blank area that sells nothing and explains nothing.
The moment visitors experience the page as finished
The largest visible element is usually the hero image or the main heading. As long as it's missing, the page looks broken or empty — even if data has long been flowing in the background. An example: a customer types your company name into her phone, the Google result appears, she taps it. From that tap, the LCP clock starts running. If your header image is in place after two seconds, she stays. If it takes six seconds, she presses back and taps the next result — your competitor. So LCP doesn't measure technology for its own sake, but exactly the wait that decides whether someone stays or leaves.
The Google thresholds: 2.5 seconds good, 4 seconds and above poor
Google divides LCP into three tiers: up to 2.5 seconds it counts as good, up to 4 seconds as needs improvement, above that as poor. These thresholds are publicly documented, and Google also uses them for its own ranking — a slow page has a harder time in search results. The audit service adopts exactly these thresholds, with no tightening and no leniency of its own. Important: measurement is done separately for mobile and desktop. A snappy value on an office computer says nothing about the phone on the mobile network, where most visitors arrive first.
How the audit service measures LCP
The audit service queries Google's PageSpeed measurement for every page checked. It supplies two views: field data from real visitors over a 28-day window, if the page has enough views, and a lab measurement under fixed, throttled conditions. Where field data is available, it counts — real visitors beat the simulation. Every value above the threshold appears as a finding in the report, with the measured value, the source of the measurement, and Google's target value alongside it. So you see not just "slow", but how slow, measured against what, and how far the way to the target value is.
- Open your homepage on a mobile phone over the mobile network, not office Wi-Fi, and watch how long the large image or main heading takes.
- Ask your agency how large the hero image file is and whether it's delivered in a modern, compressed format.
- Ask your agency to have the page's most important image load with priority, instead of queuing it behind ad scripts and fonts.
- Check the LCP value for mobile and desktop separately — the audit service's report states both independently.
- Repeat the measurement after every major change to the homepage, since a new header video can double the value without anyone noticing.
