In the latest Google Search Off The Record Podcast, John Mueller was joined by Annie and Vivek.
Annie works as a tech lead on the Chrome Web Platform Team. She works on the team that develops new core web vitals metrics.
Vivek is also on the core web vitals team, as a product manager.
Collectively, each of these extraordinary individuals offer us great insight into Google’s latest Core Web Vitals metrics.
SEO Insight #1: Page Speed is Highly Multi-Faceted
Martin posed a question about why managing speed is so difficult and why it’s such a significant topic in relation to search.
Vivek explained that speed matters a lot in Google’s own services. As they’ve worked with their own partners around the web, they’ve come to the realization that speed matters to other businesses as well.
It’s difficult for the Google team to wrap their minds around page speed, because there are so many facets that encompass it.
Because many different parts contribute to the entire feeling of page experience, there have been several long-running debates about which parts of page experience really matter, and which technologies can make a real difference.
SEO Insight #2: The Team Began With a Very Open Approach to Page Speed
Vivek details that during their questto tackle page speed, they began with a very open approach.
They did not limit themselves to any specific types of web pages – they examined as many different types of pages as they could.
They tried to understand what the common elements were that contributed to an amazing experience – the thing that causes users to go back to a website over and over again.
This is how they established the three Core Web Vitals: Largest Contentful Paint (LCP), First Input Delay (FID), and Cumulative Layout Shift (CLS).
They asked themselves about what the user really sees. What do they experience? And what are the things they encounter on a regular basis?
From the answers to these questions, they tried to derive what they wanted browser and web technologies to do.
For a long time, the answer eluded them and they didn’t realize that they could ultimately make the user experience measurable and repeatable.
SEO Insight #3: Why Do LCP, CLS, and FID Matter As Opposed to the Many Others Google Has Looked Into?
Vivek answered this question from Martin in the following way:
LCP, or Largest Contentful Paint, measures how long it takes for the user to physically see the most meaningful part of the page’s content. How quickly can they get to that content?
FID, or First Input Delay, describes how long it takes for any given page to respond to input from the user. Any interactive element of some kind plays a part in how FID is measured.
CLS, or Cumulative Layout Shift, also plays into page stability. If content on a page moves around, it can be extremely distracting and disruptive. If a page has these issues, they will factor into cumulative layout shift problems, which must be addressed and resolved.
Google really wants to ensure that any objects on the page remain where they are and where the users themselves expect them to stay.
SEO Insight #4: What Happened to Time to First Byte and the Google Speed Index?
Martin asked Annie this question, referencing his past experiences as a web developer and how people were always concerned with Time to First Byte (TTFB) and the Speed Index (SI).
Annie explained that there are certain weaknesses with these numbers.
Time to First Byte isn’t something that the user sees, it’s a technical measure of how long it took the user to receive some bytes. But this doesn’t give them information about the page itself or how it’s visible to the user..
It was a similar case with nload event, which was one of the most popular load events. The problem with it, however, was that because web pages are made out of code, they don’t really start loading until that specific onload event actually occurs.
That event metric, again, can’t even be qualified as a real user experience metric.
She was excited about the inception of Speed Index, considering it a major breakthrough, because it takes into account the average time at which pixels were physically painted on the page, and it was a better measurement of t what the user sees.
SEO Insight #5: What Does Google Try to Make Happen With Core Web Vitals?
Martin asked Annie how much of a difference that core web vitals really makes, and jokingly stated that web developers don’t need any new metrics.
Annie explained that there are two really big goals that Google has for core web vitals.
The first one is on the actual user experience, something that Time to First Byte and the Speed Index did not effectively cover.
The second goal is to be able to know exactly what the user sees when a page physically loads, not just an approximation.
The main issue with Speed Index was that it read the screen pixel by pixel.l. This brings into question many security and privacy concerns. Additionally, there are performance challenges, where they can’t exactly implement it as a performance API.
Therefore they had to find a metric that they could really use in a full user-monitoring context. Largest Contentful Paint (LCP) was the major extraordinary breakthrough in regards to that.
SEO Insight #6: Why Is There Such a Big Difference Between What is Measured
Martin asked Annie a very weighty question in this regard: Why such a big difference between tools and how they measure these things, such as differences in Chrome Dev Tools with Lighthouse vs. Web Page Test vs. Google Search Console? And compared to what can be seen in the page experience report?
There are different ways of measuring Speed Index and real user metrics. What are the advantages and disadvantages between these two?
Annie explained that the main thing about lab data is you’re basically telling the computer to process things.
The thing about lab data is that it can tell you a lot about worst-case scenarios: how does this website load in 3G or 2G? What happens if you cannot buy the absolute worst device ever on the market anymore?
Lab data should not be ignored because it can really help you dig into what’s happening physically on the page.
Lighthouse can measure things directly from the device itself, and there are going to be different results in Lighthouse for every time you run the tool. There will be a different LCP, FID, and CLS.
You will run into a distribution of users that have certain things happening. Some users will see it load really fast, and others will see it load really slow.
There’s a long tail where most people are pretty fast, and then things happen where it gets slower and slower over time.
In Core Web Vitals, Google is measuring what the 75th percentile of overall users have seen.
If three quarters of people using the website get a fast, slow time, we say that that is a pretty good result.
SEO Insight #7: What Happens When Everyone Has a Fast Connection? Will All of This Go Away?
John posed this thoughtful question to Annie. He inquired whether we still need to worry about core web vitals if everyone has high-end websites, top-tier phones, or high-performance devices in general.
Annie answered with a very interesting perspective: different types of pages, based on audience or based on content, will have different values.
You still want to check your LCP, CLS, and FID because it will tell you as your users are viewing your page whether or not your content actually is showing up quickly.
Sometimes, it’s the case that the content doesn’t show up fast.
For some sites and audiences, it’s just much easier to load swiftly. Other sites are serving different types of audiences who might have slower connections and slower devices.
These types of sites really need to make sure that they are serving those audiences. They want to see this content load quickly, regardless of their type of device.
Want to hear more? You can listen to the entire podcast here: