Skip to main content

Measuring and monitoring your page performance

Regularly measuring and auditing your page performance helps you stay aware of your changes and their impact on page speed. Consistent audits also help you focus on the right metrics and identify ways to improve overall performance.

Web Vitals

Web Vitals is a Google initiative that defines key metrics for measuring webpage performance. Among those metrics, Core Web Vitals are considered the most important.

Google collects Web Vitals data from Chrome browser sessions and stores it in the Chrome Browser Experience Report (CrUX). This dataset is publicly available, and you can access it through various tools and the CrUX API. See Google's Overview of CrUX.

Accessing CrUX data

You can view CrUX metrics in several ways:

Using the CrUX API

You can query CrUX programmatically for any domain (origin) or page URL. The API is free and offers a generous rate limit for high-volume queries. For more information, see CrUX API documentation.

Example curl command to query the CrUX API:

curl -s --request POST 'https://chromeuxreport.googleapis.com/v1/records:queryRecord?key=API_KEY' \
    --header 'Accept: application/json' \
    --header 'Content-Type: application/json' \
    --data '{"formFactor":"PHONE","origin":"https://www.example.com","metrics":["largest_contentful_paint", "experimental_time_to_first_byte"]}'

PageSpeed Insights

PageSpeed Insights is another Google tool for measuring performance. It runs synthetic (simulated) tests, meaning it does not measure actual user sessions.

Synthetic tests often report lower scores because they simulate challenging network conditions that real users may not experience.

Lighthouse

We recommend using Lighthouse for page audits. Lighthouse is an open-source tool created by Google to measure webpage quality and provide actional improvement suggestions.

You can run Lighthouse:

  • as a Chrome extension

  • through the web.dev Measure tool

  • from the command line

  • integrated into your CI/CD pipeline

Many CI/CD platforms provide plugins or workflows to automate Lighthouse tests and store results in your database.

Cached versus non-cached page performance

Testing a cached versus a non-cached page yields very different results.

Note

This is about server-side caching, not browser caching.

Arc XP's PageBuilder Engine handles caching at multiple infrastructure layers. Developers do not control when a page is cached. For more information, see Catching at Arc XP and How Does Fusion Content Caching Work.

Tip

When testing a page after making changes or deploying a new bundle:

  1. Visit the page to generate cache.

  2. Run Lighthouse tests to reflect the real-world, cached experience your users are likely to have.

Key performance metrics

Lighthouse measures many metrics. For how each metric affects your overall score, see Lighthouse Performance Scoring.

We recommend focusing on the following key metrics:

  • First Contentful Paint (FCP) - time to render the first piece of DOM content.

  • Largest Contentful Paint (LCP) - time to render the largest visible element (usually the main content). A fast LCP reassures users that the page is ready.

  • Time to Interactive (TTI) - time from page load until the page reliably responds to user input.

  • Total Blocking Time (TBT) - time between FCP and TTI where long tasks block user interaction.

Remember that third-party scripts (ads, analytics, widgets) can significantly impact these scores, and sometimes even more than your code.

Understanding what impacts your page

When building a new page, it's best to monitor Lighthouse scores from the start. Ideally, a fresh page scores close to 100, with scores decreasing as you add content and features.

When auditing an existing page:

  1. Disable third-party scripts (ads, widgets, embeds)

  2. Run Lighthouse to establish a baseline score

  3. Apply recommended optimizations to your code first

You likely have little control over third-party SDK behavior, but knowing their impact helps you make informed decisions.

PageBuilder Engine gives you a clean, fast starting point, like a blank HTML document. Third-party integrations usually have the biggest effect on degrading performance.

Setting realistic expectations

Different page types naturally perform differently. For example:

  • homepages tend to have heavier loads and lower scores

  • story detail pages are usually lighter and faster

Also, compare your site against industry benchmarks. Resources exist for performance expectations across industries (for example, e-commerce, news).

You can also run Lighthouse on competitor sites to better set performance goals.

Note

Avoid unrealistic expectations. Comparing an e-commerce homepage to a blog post detail page isn't a fair comparison.

In this section: