Core Web Vitals: What’s the point of Google’s key figures

How quickly can the user recognize content on the website? Google uses LCP to measure when the largest visible content block (in terms of visual presentation, not data volume) is displayed to the user. This can be text, images, or videos, for example. Since the largest element is usually loaded at the end and it’s often the main content element on the page, Google can get a very good impression of how long it takes for users to read or interact with the website.

With LCP, the search engine manufacturer has developed an alternative to older metrics such as load or DOMContentLoaded. These metrics tend to provide information that’s more on the theoretical end. Largest Contentful Paint, on the other hand, relates much more to the actual user experience. The very similar value First Contentful Paint (FCP), on the other hand, simply indicates how fast the first element was loaded. In many cases, however, this is completely irrelevant to the actual content of the website. Similarly, the First Meaningful Paint (FMP) metric has proven to be too complex and error-prone in the past.

As a guide, Google says that 2.5 seconds should cover everything from the first time the website is called up until the Largest Contentful Paint appears to users. It is important that this value also applies to all users. Values that you achieve as a webmaster in tests can differ greatly from the experience of different users.

The following tips can help you improve the LCP value:

• Use a framework instead of a large HTML page, so that you can better control how different elements on the page play out.
• Use a content delivery network so that a user’s location doesn’t have any influence on server access.
• Be careful when using CSS and JavaScript, making sure that the browser isn’t overloaded.
• Make images smaller to lower file sizes and reduce loading times.
• On static websites, a server cache can speed up access times.
• Build a connection to external sources ahead of time to avoid bottlenecks.

Once the website is loaded, users will want to interact with it. First Input Delay indicates the length of time between the website visitor’s first interaction with the website and the response of the website. This can be a login attempt, for example, but could also be a simple click on a link. To avoid spoiling the first impression, servers should process the input as quickly as possible. But server speed is only part of the challenge of achieving a good FID score: Webmasters should also consider users’ browsers. If the software is busy with other tasks, they cannot process the input. The most common weakness is therefore large JavaScript files that keep the browser busy for a long time.

Google states that it should take about 100 milliseconds from the moment a user triggers the interaction to the time the change is visible in the browser.

If you want to improve the First Input Delay of your website and provide visitors with a better experience, you should consider the following optimization measures:

• Try to split long tasks into smaller parts. If a JavaScript task is running, the browser cannot process other jobs. Google considers processing times of more than 50 milliseconds to be long tasks, so you should avoid them.
• Load the most important interaction elements first. This way, users can act faster and don’t have to wait for longer loading times.
• Use a so-called web worker. These processes make it possible to execute more complex tasks in the background.

Often a website is loaded in such a way that visitors can already use it before the last element appears in the browser. This has great advantages in terms of the LCP but can cause the layout to continue to shift. Content elements that jump around, however, result in users either getting confused while reading or unintentionally clicking on the wrong element.

For the CLS value, Google measures changes in the layout of a page. Every time an element unexpectedly moves about, it is recorded. Intentional changes due to user interactions are excluded here. Google not only pays attention to how many objects move and how often, but also measures how far the respective elements move. All of this is included in Cumulative Layout Shift.

For Google, a CLS value of 0.1 is considered good. If you are significantly above this, you should optimize your website. The following options can help:

• Always specify the dimensions of the built-in images in the source code of the website. This way, the browser can already reserve the corresponding space when building the layout.
• Dynamic ads that are placed on the site by external providers also mess up the layout. For this, it’s recommended to set a sufficiently large placeholder statically for the ads to avoid objects moving around.
• When trying to load fonts from the web, the layout can sometimes unexpectedly shift. That’s why it’s recommended to preload the fonts.