Google PageSpeed Insights explained — and why 100 is a trap

You might not know this, but there's a free tool from Google that gives your website a score out of 100. And nine out of 10 people are taking completely the wrong message from that tool. They'll get a 47 out of 100 and absolutely lose their Or they'll get a green 92 out of 100 and think job done, I'm going to move on. Brilliant. Both of those reactions to those two scores are completely wrong because the number with this isn't the point at all.

And if you're chasing the wrong thing, especially with this, you're wasting your time, you're wasting your money, and the actual problem with your website still sits there quietly costing you customers. And here's what we're going to go through with it today. What this tool actually is, what those terrifying sounding numbers and words actually mean in plain English, and why a perfect score on the tool is a complete waste of your time. But the thing here with all of this that surprises most people is actually the thing that you don't really expect. Yes, that's clickbait.

Let's get on with it. So right, we're going to have to get into first of all what PageSpeed Insights actually is because the name PageSpeed Insights actually sounds a lot more technical than it actually is. It's not actually as scary as it seems, even though I'm using the word actually quite a lot. Google PageSpeed Insights, to give it its full name, is a tool that is free to use that sits on the internet at pagespeed.web.dev. You paste your URL into it and it spits a number out to you.

That number is a score out of 0 to 100. Generally, the way that the score breaks down is that if you get a red number, which is 0 to 49, your website is bad and it needs some serious fixes. If you get a 50 to 89 score, it'll give you a little orange circle around the number. Your website needs work, but it's not the end of the world. And what we're aiming for here is a green circle, and that score will be somewhere between 90 and 100 out of 100.

That means you've got a good website, there's probably nothing else that you really need to fix, and Google PageSpeed Insights is happy, and you should be happy, too. Job done. I'm doing thumbs up a lot in these videos at the minute, aren't I? We need We should bring the thumbs up back. Google PageSpeed Insights tests mobile and desktop separately, and it'll give you two separate scores.

So, when you flick between mobile and desktop, you might notice that you've got, I don't know, 50 out of 100 on mobile, but 90 out of 100 on desktop. But really, the mobile score is the one that really matters, and that's the one that's usually going to be the worst score out of the two. And by the way, if your desktop score is lower than your mobile score, there's probably something very wrong with your website because it should be the other way around. It's generally much easier to get a high desktop score, and it's not the one that we really care about because more than likely the majority of your users are coming to your website using a mobile device. On most of the websites that we look after these days, we get probably about 80% traffic from mobile, so we can pretty much disregard the desktop scores.

Not that it really matters here, but a bit of technical mumbo jumbo, the actual Google PageSpeed Insights tool is a tool called Lighthouse, and you can use that tool directly inside Google Chrome browser. It's All that Google PageSpeed Insights is really doing is using that tool, but online. And then on top of that, I said in a little bit of extra mumbo jumbo to make it even more confusing for you. In general though, the only thing you need to know about Google PageSpeed Insights is essentially a free health check for your website. You punch your URL in, and it'll give you a number back.

Now though, we need to talk about what those numbers and those scary words actually mean. So, once you've run your website through Google PageSpeed Insights, you're going to be presented with a page that looks a little bit like this, and you're not really going to know what you're looking at, and it's going to look a little bit scary. They're going to present things to you called Core Web Vitals, which sounds like an hospital drama or something, but don't worry. I'm going to actually talk about what each individual one actually means. There's that word again there, actually.

Actually. Mainly though, what PageSpeed Insights is actually doing is just giving you an opportunity to answer a couple of very simple questions. How long until I see something? How long until the biggest thing shows up on the page? Is the website frozen, and if so, for how long?

How quickly does the page look finished? And does stuff jump around on the page when it's loading? Your number out of 100 is also here, right at the top, and that's what we talked about earlier. We're currently getting orange on the Genius Division website, which means we're three points away from figuring out how to make it good. Another thing to point out as well is that PageSpeed Insights can go up and down, so you should check it pretty regularly.

On all our websites that we look after, we check this score every single day on mobile and desktop, and plot it on a graph over a month because they can go up and down. Let's deal with the first score first, FCP, our first contentful paint. That's this bit here. We're getting 1.1 seconds for first constant Can't say the word. FCP we'll call it from now on.

And this first score, FCP, is basically saying, "How long until I see something on the page?" It doesn't matter what that is. When somebody opens your website, this is measuring how long it takes for the first thing to appear on the screen. Think of this like um if I had a coloring book, for example. Before the website loads, the coloring book is completely closed. When the coloring book opens, how long does it take before the first thing is visible in my eyes?

That's probably a bit of a complicated metaphor, actually, isn't it? For us here, you can see that we're getting 1.1 seconds, which is actually a pretty damn good figure, and visitors aren't staring at a blank page for a very long amount of time. The next one we should look at here is the largest contentful paint. It's often called LCP. This is generally trying to measure how long it takes for the most important thing to load on the page.

Basically, the idea that the most important thing is probably going to be the biggest thing on the page. So, it's considering things like the hero image at the top of the site or the main headline at the top of the page, the one that loads in first, or a large video thumbnail, all those kind of things. Think of FCP as your waiter says to you, "Your pizza is on its way." And then LCP is when the pizza actually arrives. Again, that's another quite confusing metaphor, really, isn't it? On our page at the minute, the LCP is actually a little bit slow to load, and we probably want to look at that.

In general for LCP, Google wants under 2.5 seconds. They consider that really good. 2.5 to 4 seconds is considered to need improvement, and then over 4 seconds, you'll get a red, and they consider that to be very slow, and it definitely needs attention. Now, this one's a weird one, really, because your LCP, your largest contentful paint, could change fairly regularly. If you decide to load up a new huge image on your home page, or whatever page you're currently testing right now, that could become the new LCP.

And if you take that image off, it won't be that anymore, it'll be something else. So, I think the LCP, probably out of all of the ones, is the one that's the most complicated to explain, and also the one that's the most complicated to understand. The next one I want to talk about is total blocking time. This is usually called TBT, because for whatever reason, PageSpeed Insights really loves an abbreviation. This one here, TBT, is basically referring to the question, "Is the website frozen?" And if it is, how long does it take before the website becomes unfrozen?

Sometimes websites get busy when you're trying to load them, or sometimes they just stop responding. And that's what this is testing. How long does it take for the website to go from not responding to responding? Ideally for this we do want TBT, our total blocking time, to be zero. We don't want our website to be thinking.

Every second that our website is not responding is another second for our customer to disappear. The next one I want to talk about is speed index. So, that's the bit lower down. Uh where I'm getting 1.2 seconds on this. And the question that this is is how quickly does the page look finished?

So, this is measuring how fast the screen fills in when the tool is testing it. The way to think about this in a web browser context is that uh when a web browser loads a website, it is drawing a picture. And as it slowly draws the picture in, the website becomes revealed. And what speed index is doing is measuring how long it takes for the full picture of the website to be drawn into the web browser. Final one I want to talk about is CLS, which is cumulative layout shift, which is actually very difficult to say.

This is answering the question, does stuff jump around on the page? And on most web pages, things do jump around quite a lot because we're often using custom fonts now. Uh we're not putting image dimensions on our images. And it means that when a web page loads in and you start scrolling it before it's fully loaded. By the way, news websites are terrible for this.

As you scroll down the website and more and more of the website loads, things start moving around. And you'll lose the spot where you were currently scrolling to. So, CLS does want to be zero or as low down as we can get it. This, from a user point of view, is probably one of the most frustrating things that you definitely need to fix. Because what Google PageSpeed is testing here is how much the page moves around.

And we don't want that to happen for our users. Overall here, you can see with my figure 0.17, it's actually red and it's not very good. Something that we need to look at fixing. I should have probably fixed it before I made this video about PageSpeed Insights. And that's the whole thing in plain English that you really need to understand with page speed.

But the big myth here that's been hidden is under that number that we looked at before, and that's really the most surprising thing. And yes, I am being purposefully clickbaity again. And we're going to get right to what I mean by that in a second after the show break. No. No, sorry.

We're not having a show break. I'm just joking. Let's move on. So, the most surprising thing I'm getting at here is the thing that most people get wrong. And it's the most important thing I want you to take out of the entire video.

And it's the number. Everyone treats the number out of 100 like it's an exam and you want full marks. Everybody wants to get 100 out of 100. Everybody wants to get as close to that as possible. When in truth, once the number has gone green, you don't need to spend any more effort on it.

You should put it to the side and leave it. You should monitor it and keep checking it, but don't try and get from 94 to 96 or 96 to 98 because you'll be spending a lot of wasted time. You'll make no more money in your business. And what you're really doing is not improving your website on the things that should matter like fixing your content, improving your images, making sure you've got a pricing page, those kind of things. So, this can be a bit of a rabbit hole that I don't suggest you go right down.

So, if you don't need 100, why am I even telling you to care about this stuff at all? There is actually two very good reasons about why you should care about this. And number one is that people are going to leave your website. If your website is slow, it will be bleeding visitors. So, if we've got a bad TBT, total blocking time, if it takes ages for the website to become responsive, we're going to lose people.

If the website takes 5 seconds to load, if it's 5 seconds on a speed index, we're going to be losing people. If the biggest thing on the page, the LGS, the largest contentful paint, is huge and takes ages to load, we're going to lose visitors. So, the first thing to remember is that users will leave if your score is too low. This directly affects your users' experience of your website. The other thing to say, number two, is that Google does use Google PageSpeed as a ranking factor.

The speed of your website is a specific ranking factor that they're taking into account when they're ranking your website. And they also use your figure, your score out of 100, to rank your website. They won't tell you specifics about how this actually works, but we do know, and we have been told by Google, that this is a specific ranking factor. Therefore, if it is, we should pay attention to the advice, and we should make sure that our score is as close to 100 or 90 out of 100 as we can get it. So, if we have a slow website, it's a double whammy.

We're getting less visitors because they're leaving because they're pissed off the site is slow, and we're also ranking worse because Google is paying attention to Google PageSpeed, and then ranking the website lower. So, that's Google PageSpeed Insights. It's a free health check. It's not an exam. You're aiming for green, and not 100, usually.

And when your score is really bad, and I will probably stake money on this, it's usually the images on your website. So, if you've never run it before, I suggest you go over to PageSpeed Insights website now. The link is in the description. Paste your URL in, hit the mobile tab, and look to see what's in it. And yes, a fast website is brilliant, but if the people who land on your fast website can't figure out what you're doing in about 4 seconds, speed doesn't even matter.

And that's exactly what I get into into this video. Five questions every customer will ask about your website homepage. Go on. Go watch it.