PageSpeed.ONE Methodology
Awards 2026

Measuring website speed seems simple: open a page and count seconds. But speeds vary between a developer on fiber in an office and someone on a train with an older phone. That's why Awards rankings are based on real user experiences, not one-time tests. Here's how the numbers are generated and interpreted.

Why We Use Real User Data Instead of One-Time Tests

Instead of a one-time bot, we use data from Google's public database, the Chrome UX Report. Chrome anonymously records how quickly websites loaded for real users over the 28 days before the measurement on October 19–20, 2026.

The results reflect real user experience, not just lab conditions. We explain the difference in our article on Comparing Lab Tests and Real User Data. Discover why website speed matters in our overview Why Website Speed Matters.

One Number Instead of a Sea of Acronyms: PageSpeed.ONE Score (SPS)

Tracking six different technical metrics separately for mobile and desktop can be confusing. That's why we created the PageSpeed.ONE Score (SPS) on a scale of 0 to 100.

Just like in the app: 76 points and above (green) means your site is performing well, 26 to 75 points (orange) indicates room for improvement, and 25 points or less (red) significantly slows down your site.

If two websites have the same overall score, the public significance of the institution decides (e.g., a ministry over a smaller office).

Websites are ranked by overall SPS in the table. We display scores for both mobile and desktop. The Speed Test link initiates a live measurement on PageSpeed.ONE.

Three Smoothness Tests (Core Web Vitals)

The score is based on Google's three Core Web Vitals. To earn a Pass label, a site must excel in all three for at least 75% of visitors. We evaluate the 75th percentile, not the average, to avoid masking poor experiences for a quarter of users.

• Loading Speed (LCP): How quickly does the main content appear? To avoid a blank stare, it should be within 2.5 seconds.
• Interaction Speed (INP): How quickly does the site respond when you click a menu, filter, or button? A good site responds in under 200 milliseconds. Longer delays make users feel the page is frozen, prompting repeated clicks.
• Visual Stability (CLS): Does the page jump around while loading? You know the drill: you aim for a link, but a banner loads, shifting the text, and you miss. A stable site keeps the shift score below 0.1.

Time Window and Data Collection

Values are calculated from a 28-day window before the measurement on October 19–20, 2026. The result reflects the long-term health of the entire website, not just a single outage on Tuesday afternoon.

For a technical overview of how PageSpeed.ONE measures and collects data, see the documentation How We Test and Source Data. You can check your own site with the Website Speed Test, and for daily monitoring, use Monitoring PLUS.

Why Some Websites Are Missing from the Table

Google only publishes statistics for websites with sufficient traffic to ensure data reliability and prevent individual identification. Therefore, a small community or new political party with lower traffic might not appear in the interactive ranking. However, they can be found in the open data (CSV download), just without measured values.

Government and Political Parties: Source of the List

The preliminary ranking of government and political parties is based on the list of institutions and categories from index.gov.cz and index.csgov.cz. Thanks to both project teams for the open data. We added political parties and standardized redirects for some domains (e.g., to gov.cz). This is not an official state list, but a guide for orientation.

Open Data, License, and Contact

We publish ranking data under the CC BY 4.0 license. Please credit PageSpeed.ONE Awards and link to the relevant ranking page. Questions or found an error? Contact us via the contact page.