We delved into 208,085 web pages to better understand Core Web Vitals.
First, we set benchmarks for Cumulative Layout Shift, First Input Delay, and Largest Contentful Paint.
Next, we explored the connection between Core Web Vitals and user experience metrics, such as bounce rate.
WebCEO provided the data, which revealed some intriguing findings.
Let’s examine the data further.
Optimal Largest Contentful Paint Scores Found in 53.77% of Websites
Our initial objective was to evaluate each site’s performance based on Google’s Core Web Vitals: Largest Contentful Paint, Cumulative Layout Shift, and First Input Delay. These vitals contribute to Google’s comprehensive assessment of “page experience.”
Specifically, we aimed to identify the percentage of pages categorized as “good,” “needs improvement,” and “poor” within each site’s Search Console.
We scrutinized anonymized Google Search Console data from approximately 208k pages (around 20k total sites) to accomplish this.
Our first assignment is to analyze LCP (Largest Contentful Paint), which measures the time it takes for a page to load its visible content.
The outcomes of our analysis were as follows:
53.77% of websites had an optimal largest contentful paint score
- Good: 53.77%
- Needs Improvement: 28.76%
- Poor: 17.47%
As evident, most of the sites we assessed had a “good” LCP rating, which was higher than anticipated, particularly considering other benchmarking efforts (such as iProspect’s).
Our dataset’s websites may be particularly attentive to page performance, or the difference in sample size may have played a role (iProspect’s analysis monitors 1,500 sites continuously, while we examined over 20,000).
Regardless, it’s reassuring that only about half of all websites need to improve their LCP.
53.85% of Analyzed Websites Had Good First Input Delay Ratings
Next, we investigated First Input Delay (FID) ratings reported by Search Console. FID measures the delay between the initial request and a user being able to input something (e.g., entering a username).
Here’s a breakdown of FID scores from our dataset:
53.85% of analyzed websites had good first input delay ratings
- Good: 53.85%
- Needs Improvement: 37.58%
- Poor: 8.57%
Once again, over half of the sites, we examined had “good” FID ratings.
Interestingly, very few (8.57%) had “poor” scores, indicating that a relatively small number of sites are likely to be adversely affected when Google integrates FID into their algorithm.
65.13% of Sites Exhibited an Optimal Cumulative Layout Shift Score
Lastly, we reviewed the Cumulative Layout Shift (CLS) ratings from Search Console.
CLS measures the movement of elements on a page while loading. Pages that remain relatively stable during loading have high (good) CLS scores.
The CLS ratings among the sites we analyzed were as follows:
65.13% of sites exhibited an optimal cumulative layout shift score
- Good: 65.13%
- Needs Improvement: 17.03%
- Poor: 17.84%
Of the three Core Web Vitals scores, CLS was generally the least problematic. Around 35% of the sites we analyzed need to improve their CLS.
Average LCP Stands at 2,836 Milliseconds
Next, we aimed to establish benchmarks for each Core Web Vital metric. As mentioned earlier, Google has developed its guidelines for each Core Web Vital.
(For instance, a “good” LCP is considered to be under 2.5 seconds.)
However, we hadn’t encountered any large-scale analyses that sought to benchmark each Core Web Vital metric “in the wild.”
First, we benchmarked LCP scores for the sites in our database.
The average LCP was 2,836 milliseconds (2.8 seconds) among the sites we analyzed.
The most common issues that negatively affected LCP performance included:
Issues affecting LCP
- High request counts and large transfer sizes (100% of pages)
- High network round-trip time (100% of pages)
- Critical request chains (98.9% of pages)
- High initial server response time (57.4% of pages)
- Images not served in next-gen format (44.6% of pages)
Overall, 100% of pages had high LCP scores, at least partially due to “High request counts and large transfer sizes,” which means pages that are heavy with excess code, large file sizes, or both.
This finding aligns with another analysis we conducted, which discovered that large pages were primarily responsible for slow-loading pages.
Average FID Measures 137.4 Milliseconds
We then examined FID scores among the pages in our dataset.
In general, the mean First Input Delay was 137.4 milliseconds.
The most prevalent FID-related issues we identified included:
Issues affecting FID
- Inefficient cache policy (87.4% of pages)
- Long main-thread tasks (78.4% of pages)
- Unused JavaScript (54.1% of pages)
- Unused CSS (38.7% of pages)
- Excessive Document Object Model size (22.3% of pages)
It was intriguing to see that caching issues tended to have a more negative impact on FID than any other problem. And, as expected, poorly-optimized code (in the form of unused JS and CSS) contributed to many high FID scores.
Average CLS Equals .14
We discovered that the average CLS score is .14.
This metric specifically examines the “shifts” in a page’s content. Anything below .1 is rated as “good” in Search Console.
The most common issues affecting projects’ CLS were as follows:
Issues affecting CLS
- Large layout shifts (94.5% of pages)
- Render-blocking resources (86.3% of pages)
- Text is hidden during web font load (82.6% of pages)
- Not preloaded key requests (26.7% of pages)
- Improperly sized images (24.7% of pages)
How LCP Relates to User Behavior
Now that we set benchmarks, we wanted to determine how accurately Core Web Vitals reflect real-life user experience.
Google emphasizes this relationship in its “Core Web Vitals report” documentation.
To examine the impact of Core Web Vitals on UX, we looked at three UX metrics designed to represent user behavior on web pages:
- Bounce rate (% of users leaving a website’s page upon visiting it)
- Page depth per session (how many pages users view before leaving the website)
- Time on the website (how long users spend on a website in a single session)
Our hypothesis was: improving a website’s Core Web Vitals would positively affect UX metrics.
In other words, a site with “good” Core Web Vitals would have a lower bounce rate, longer sessions, and higher page views. Fortunately, besides Search Console data, this dataset also contained UX metrics from Google Analytics.
We then compared each website’s Core Web Vitals against each UX metric. The results for LCP are shown below:
In the three graphs, it was evident that all three segments (Good, Poor, and Needs Improvement) were evenly distributed.
In other words, there wasn’t any direct relationship between LCP and UX metrics.
FID Shows a Mild Connection With Page Views
Next, we investigated the potential relationship between First Input Delay and user behavior.
As with LCP, it’s reasonable to assume that a poor FID would negatively impact UX metrics (especially bounce rate).
Users who have to wait to select from a menu or type in their password will likely become frustrated and bounce. And if that experience persists across multiple pages, it may reduce total page views.
With that in mind, here’s how FID correlated with the behavioral metrics.
Note: We found that a high FID correlates with a low number of pages per session. The opposite was also true.
FID and Time on Site
In general, we only observed hints of correlation when comparing FID to the number of pages viewed per session. A website’s FID does not influence user behavior regarding bounce rate and time on site.
How CLS Affects User Behavior
Next, we wanted to explore a potential link between CLS and user activity.
It seems plausible that a poor CLS would frustrate users and, as a result, increase the bounce rate and reduce session time.
However, we need help finding case studies or large-scale analyses demonstrating high CLS scores influencing user behavior. So, we decided to analyze for potential relationships between CLS, bounce rate, “dwell time,” and pages viewed.
In summary, we didn’t observe any significant correlation between CLS, bounce rate, time on site, or page views.
Conclusion
I hope you found this analysis interesting and helpful (particularly with Google’s Page Experience update approaching).
I want to thank SEO software WebCEO for providing the data that made this industry study possible.
It was fascinating that most of the sites we analyzed performed relatively well and are largely prepared for the Google update. It was also interesting to discover that, while Core Web Vitals represent metrics for a positive UX on a website, we didn’t observe any correlation with behavioral metrics.
See also: 6 Essential Questions You Need to Know on How to Achieve Keyword Research for Effective SEO
For SEO audits, click here to book your appointment.
What’s Your SEO Score?
Enter the URL of any landing page or blog article and see how optimized it is for one keyword or phrase.
