Earlier this year, Google released Lighthouse 6.0, the latest version of its automated website auditing tool, which landed in Chrome this July. The new version includes a re-weighting of metric priorities and adds new criteria – all in service of emphasizing main thread interactivity and layout predictability.
The latest Lighthouse version includes the following scoring changes:
- First Contentful Paint was reduced to 15% in Lighthouse 6 compared to 20% in 5.
- Speed Index was reduced to 15% from the previous 27%.
- Time to Interactive was reduced to 15% from the previous 33%.
- A new metric, Largest Contentful Paint, is weighted at 25%.
- Two other new metrics are also included in version 6 – Total Blocking Time (25%) and Cumulative Layout Shift (5%).
- In version 5 the metric First Meaningful Paint was weighted at 7% but has been removed from version 6 because it was deemed to be too variant and not standardizable.
- Version 5 metric – First CPU Idle (13%) – was not included in version 6. Max Potential FID was also deprecated in version 6 but is still available in the JSON.
In the new scoring, Largest Contentful Paint (LCP) and Total Blocking Time (TBT) took over as the two most heavily weighted metrics. LCP represents the time between a user starting a page load and the page rendering its primary content and measures when the user perceives the largest content of a page is visible. TBT measures the total amount of time a page is blocked from responding to user input such as mouse clicks, keyboard activity and screen taps.
|First Contentful Paint||15%|
|Largest Contentful Paint||25%|
|Time to Interactive||15%|
|Total Blocking Time||25%|
|Cumulative Layout Shift||5%|
|First Contentful Paint||20%|
|Largest Meaningful Paint||7%|
|Time to Interactive||33%|
|First CPU Idle||13%|
The addition of a scoring calculator, provides a great tool to compare how your website performance across both Lighthouse versions. And after completing an audit, the calculator provides a dynamic way to assess how each audit value directly impacts your overall score.
One thing to keep in mind with any lab measurement tool, such as Lighthouse, is the tools are easy to use and provide a valuable snapshot of site performance, but lab tools are limited in that they don’t necessarily reflect real performance.
This issue arises in several ways, such as basing a score on a cold load with an empty cache rather than a user experience with a warm cache. Analytics tools often only score the first-page load and ignore the speed improvements over a browsing session from single page applications (SPAs). Outside of the lab, prefetching – particularly predictive prefetching combined with edge computing where websites stream dynamic content to the browser before it’s requested – delivers a sub-second experience, but is penalized in Lighthouse. And finally, Google prerendering AMP on mobile SERPs leads to an instantaneous page loads, yet isn’t factored in performance tools.
Even with its limitations, Lighthouse is a valuable barometer for performance.
Google never stops experimenting, and future versions of Lighthouse are sure to come online. In fact, Google is already testing with collecting source maps to power features, like:
- Spotting excessive polyfills or transforms in code sent to browsers.
- Treemap visualizations to highlight modules that require action.
With the introduction of Core Web Vitals, and the Chrome team’s push for a more holistic view of a user’s experience, speed and performance are bedrock attributes of any website. As you push for a faster mobile experience, understanding how to glean actionable insights from Lighthouse, helps inform what enhancements to pursue.
If you have any questions about how to use Lighthouse, or what techniques can tick all the speed boxes, please set up some time to chat.