Core Web Vitals for Enterprise Businesses: Q/A with our own Kathy Brown and Karl Kleinschmidt
On February 18th, we hosted our Core Web Vitals for Enterprise Businesses Session to address Google’s new Page Experience ranking factor update. The session focussed on what the Core Web Vitals were and why they mattered. Our experts dived into what each of the new ranking signals meant and how they would impact SEO. Although we provided actionable tips and advice, there were still many questions around FID, LCP, CLS, and more.
Below are the questions collected from our audience and our experts Kathy and Karl address them.
What is the difference between Field and Lab data?
Kathy: Lab data are the performance data you see in a specific environment. Tools such as Chrome Dev Tools, as well as tools like webpagetest.org are providing you Lab Data. Field Data is data collected from the many users that are using Chrome to browse your site’s pages. You can see Field Data in Google Search Console and often in Google Page Speed Insights (which will report both Lab and Field data for a page). For automation needs, Field data are accessible via BigQuery. Remember that Lab data is for testing, Field Data is for ranking. Based on comments from John Mueller, we are anticipating that initially, CWV will only impact mobile rankings and that you will need to be “in the green” for all three metrics to rank higher.
What do you do when there are no Field data available?
Kathy: If there are no Field data available this likely means the page doesn’t get a lot of traffic. Use Google Page Speed Insights to check your most popular pages to see if Field data are available for those pages. You can then extrapolate the findings for all your pages that belong to that page type. Try setting up a CrUX report to see if there is one for your origin (your domain). You can find a prebuilt Google Data Studio template and this will give you performance data for your site overall.
If you still can’t find any Field data, then determine the top most common devices and browsers that are accessing your site and test using those environments. Remember to set the throttling and bandwidth controls to simulate as close as possible your visitors’ environments.
During development, our websites are run on secondary servers which are way slower than our production servers. How do you go about testing page speed in this situation?
Kathy: A significant amount of Core Web Vital scoring will be dependent on the client and the code that it is running. So a poor CLS score has nothing to do with a slow server. However, LCP could be impacted by a slow server due to slower connection handshakes and the time it takes to get resources sent over. One approach to consider is to create a comparison for your servers for metrics such as TTFB (time to first byte) for a variety of pages, as well as comparing the time to deliver each 1 KB of data. Additionally, if you have database operations or server-side JS, knowing the differences between those execution times in each server would also be helpful. Compare the waterfall charts between the two and see how much additional wait time there is in the development environment. Knowing these differences would help evaluate the LCP performance on your development server more easily.
Do you think preloading links (for subsequent pages) will impact LCP OR do you think LCP is dependent only on the first page that the user loads?
Kathy: When it comes to Field data, the performance of all pages will matter. Caching static assets (which helps for subsequent visits and pages) will help your Field scores. Preloading CSS styles using the link, preload will likely also help the performance of all pages.
Could a CDN with good cache optimization significantly decrease the LCP score?
Kathy: It is certainly possible if the LCP is appearing more quickly for all users (both new and returning) then your LCP score will improve.
How do you use popups while not having them affect Core Web Vitals?
Karl: For LCP issues, verify that popups don’t take up too large of a percentage of the viewport, especially on mobile.
For CLS issues, verify that popups are not making the rest of your elements move, but are in front of them.
How should we be thinking about CWV in the context of mobile-first indexing?
Karl: CWV is going to be rolled out to mobile-first and delayed for desktop (we don’t know by how much). CWV is not really about how Google crawls your site, but instead how visitors interact with your site, so mobile-first indexing shouldn’t make a difference.
What’s the difference between Total Blocking Time (TBT) and FID?
Karl: Total Blocking Time is the total amount of milliseconds main thread blocking tasks is blocking interaction (100 ms are subtracted per task). FID is the average of how long it takes for your website to respond to visitors’ interactions. Total blocking time is therefore the total amount of milliseconds in the loading of the page where you could have an FID above 0, while FID captures how users are actually interacting with your site.
We see score changes in Google Search Console without us making changes/improvements on our end.
What would be the reason for these fluctuations within the search console?
Karl: For Search Console is field data, it depends on how users interact with your site. Changes in user behavior can mean different CLS scores because visitors are scrolling differently or different LCP scores because a higher percent of visitors are returning customers so they have certain images cached. There are many possible causes for the fluctuations in metrics and remember the data is delayed by around 28 days, so you might have released something 28 days ago and are now seeing the changes in data.
If someone is getting different results from Search Console and PageSpeedTest, should we only worry about what is reported on Search Console?
Karl: Outside of Field vs. Lab data, there are two main causes for different data between search console vs. page speed insights. If one of your pages hasn’t received enough traffic, Google uses similar page data in page speed insights, while in search console is grouped similar pages that could have enough data. It is therefore also possible that one of your pages has enough traffic, but none of the similar pages have enough traffic, which could also cause a difference in scores. There was also an update to search console reporting, which could have had an impact.
Need additional resources?
We have a lot of great content on this topic. Check out our Core Web Vitals page where we’ve assembled a library of expert resources like podcasts, webinars, and blogs. You can find everything you need to get ready for Google’s Page Experience ranking factor update.