What is Core Web Vitals?
Core Web Vitals are three metrics that quantify a user’s experience while browsing a webpage. Content loading speed, interactivity, and visual stability are all hallmarks of the user experience (UX).
Since mid-June 2021, Google has been using these as SEO ranking factors. These three metrics are the LCP (Largest Contentful Paint), FID (First Input Delay) and CLS (Largest Contentful Paint).
Largest Contentful Paint
LCP metric measures loading performance. To provide a good user experience, LCP should occur within 2.5 seconds of when the page first starts loading.
First Input Delay
FID metric measures interactivity. To provide a good user experience, pages should have an FID of 100 milliseconds or less.
Cumulative Layout Shift
CLS metric measures visual stability. To provide a good user experience, pages should maintain a CLS of 0.1. or less.
Most page performance testing tools will test these 3 metrics also. However, those are usually lab tests rather than anonymized actual user page experiences.
What is Chrome User Experience Report (CrUX)?
Under Google Search Console, the Core Web Vitals report’s data comes from the Chrome User Experience Report (CrUX). This reflects actual usage data on your site from users around the world.
The CrUX report gathers anonymized metrics about performance times from actual users visiting your URL. Hence it is much more accurate to the actual realistic page experience of users on your web page.
When your website has enough traffic and is captured in the CrUX. You can see it in the Page Experience report under the Google Search Console showing the approximate percentage of Good URLs in Core Web Vitals performance and how many total impressions they have. Numbers are being estimated since source reports calculate values over different time frames.
Generally, you can see an overview of how good your website’s page experience is and how many times those pages show up in Google Search Engine.
The Challenge
Previously the client website is manually optimized by different plugins like LiteSpeed Cache, Dynamic CSS, Dynamic JavaScript, and Minifying, along with Deferment of jQuery and so on. Images are also uploaded with a suitable resolution and along with image compression plugins, and lazy loading of below-the-fold images during page load.
The website page speed is good, and the foundation is solid to withstand most of Google’s algorithm updates.
The Page Experience is quite good and stable, with an average of 95% Good URLs and an average of 4,500 daily impressions of those Good URLs.
Even after the Desktop’s Page Experience Update rolled out in February 2022 making it a ranking factor, the search performance for Desktop devices wasn’t impacted much.
NitroPack Controversial Optimization
The client decided to disable all plugins and optimization and opt for NitroPack paid plugin. The said solution had some controversy on the method being used, and there was some issue that will arise if it was not properly configured and managed.
The configuration has been set to the highest level of optimization. We first notice it when doing a checking of the website, the actual user experience we had was subpar compared to the previous optimization, especially on the Desktop view.
Can see that the image’s lazy load is slow, only top-of-the-fold assets are being loaded as static, with no user interaction being recognized. After around 3 seconds only the page flashes and loads again with proper assets and allows the user to click or scroll.
However, most testing tools return very good results. Although we do see that it seems to be using a controversial method and focus on getting a high score from most of the testing tools. While actual user experience is comparatively worst.
Poor Google Page Experience
Due to the way most testing tools are based on the lab test result, NitroPack can recognize and specifically optimize those testing tools. The result is always great, hence there’s no solid proof to see that the real Page Experience has worsened.
The only option is to monitor, and then check the Page Experience Overview report which is based on the real website’s user experience from Chrome User Experience Report (CrUX). The reports are evaluated across a time frame, hence they will be estimates and averages, and only usually able to see the effect after 2 weeks to 1 month with enough traffic and data from real users.
Each CrUX Report data point collected on any day represents an aggregate look at a trailing 28-day period. The aggregation for performance is reported at the 75th percentile. For example, a CrUX Report data point created on March 29 would include data from March 1 through March 28 and give you the 75th percentile of performance for all users in the data set.
Ref: What Google’s Chrome User Experience (CrUX) Report is, and Why You Should Care (bluetriangle.com)
After around 4 weeks, can see from the Desktop Page Experience Overview that Good URLs started dropping from an average of 95% to 13.7% lowest point, and from an average of 4,500 impressions daily to 242 impressions only at the same lowest point. Mobile Device Page Experience Overview Good URLs are not being affected but can see an obvious slight drop in the impressions count.
Impact on Search Results Performance and Ranking
By comparing the performance of search results before and after NitroPack was implemented (before our suggested solution). The image below shows the impact of poor Desktop Page Experience on clicks and impressions. There are minor ranking changes also, but is an average drop by 1 position.
Although there might be other issues contributing to the difference in the clicks and impressions of this Desktop device’s Search Results Performance.
But in this case, it is quite unlikely, due to the pattern of lower search results performance on the Desktop matched with the timing of the poor Desktop Page Experience
Our Approach to this SEO issue
We already did our research on said SaaS optimization tool and notice that any negative effect will show in the CrUX report within a month. Therefore, we recorded the date of the changes being made and alerted the client on possible issues if the configuration isn’t right.
After a month, there’s already proof of the Desktop device’s Page Experience is bad which can also see from the Search Results Performance report.
1st solution was for the client’s agency to check with NitroPack and submitted sitemaps to make sure desktop pages will be crawled. Something that doesn’t align with us is that the Mobile device’s page experience is not being affected too much. However, all parties decided to let the NitroPack side do their things first.
However, after another month, the improvement is still not being seen, and the Page Experience at Google Search Console is still bad for Desktop devices.
Root Cause is at Page Experience Signal
Other than researching on NitroPack, we also dug into the CrUX report. We have compared the actual user data from before and after the change.
It is very obvious that there’s an increase in Poor CLS from 3.2% to 18.57%, and those that need improvements also increased from 5.96% to 14.07%. Indicating that there’s a drop of Good CLS from 90.84% to just merely 67.37%.
Solution (Core Web Vitals metrics are important)
When we bought this up to the client, they approached their vendor to try to solve this. The tests being done by the vendor are all still lab tests, trying to prove that nothing is wrong. Hence, we pointed them to the Page Experience Overview still shows a downward trend of Good URLs and theirs’s impressions.
We also suggested that it may be the issue of the optimization configuration has been cranked up to the max, causing these issues.
The vendor then request support from NitroPack and got a response that it was due to the Desktop view of those pages not being optimized or crawled by NitroPack, and they requested to submit the Sitemap to make sure all pages are being optimized.
Something that doesn’t align with us is that the Mobile device’s page experience is not being affected too much. However, all parties are on board with the suggested solution from NitroPack since they might encounter these scenarios before and know what’s best.
After another month of monitoring, the Desktop Page Experience is not trending up at all. We follow up with the client’s vendor, and also discuss our findings, and use a reporting tool to show them the comparison of the month-to-month CrUX report, especially on the CLS issue.
The client’s vendor finally agrees to drop down a notch for the optimization configuration level.
Result
Just after about a week, Page Experience has shown an obvious increase in Good URLs and also daily impressions count trending upward. We also notice an improvement in the daily impression at the Mobile Page Experience Overview.
After about a month, we can see from the Google Search Console that Desktop Page Experience is back to normal before NitroPack’s max-level optimization. Then we look at the Search Results Performance and can see that the performance has been significantly improved.
Definitely needed more time to catch up to the level
Recommendation
Would recommend that when utilizing any website optimization tools or plugins, have to test the website’s actual user experience. Cannot always rely fully on only some testing tools that usually are considered lab tests. The best is to get a few users on different devices and connections to browse through the website naturally.
Those sites with high traffic and have data for Page Experience Overview and CrUX Report can keep actively monitoring the Page Experience Overview for 1 to 2 weeks later.
If need any help to take a quick look at your website, can drop us a message anytime.
References:
NitroPack Review – Cheating Tests w/Blackhat Techniques? (wpspeedfix.com)
Is NitroPack plugin Black Hat SEO for speed optimization? – Google Search Central Community
