Old vs. New Google PageSpeed Insights – Putting it to the Test

Page speed is more important than ever.

It’s built into Google’s search algorithm, so a speedier competing website could outrank yours and leave you in the dust! It’s unavoidable, and it applies to everyone. Version 5 of Google PageSpeed Insights makes speed on mobile ever more crucial. Check your pace and start planning for your SEO efforts ahead.

If you’ve been testing the speed of your website lately between sprinting to catch up with your competitors, you may have noticed something has changed.

Update May 2020: A set of Core Web Vitals were announced by Google as signals they believe to be essential to delivering a great user experience, with plans to incorporate these into an Page Experience algorithm update sometime in 2021. Pagespeed plays an important part; one of the Core Web Vitals announced was the loading time of a page – specifically, the Largest Contentful Paint (LCP). The PageSpeed Insights tool was updated once again, this time to include reports on the Core Web Vitals in both lab and field data.

Read more: Why Is Pagespeed Important?

Jump to:

• Page speed, redesigned
• Old vs. New Google PageSpeed Insights
• What’s changed in the results?
• Does speed really matter?
• Putting Google PageSpeed Insights to the test
• Choosing the test subjects – American Airlines
• Choosing the test subjects – Tyson Gay
• Comparing the results
• The implications
• The need for speed

Page speed, redesigned

Google has rolled out a new version, v5, of their PageSpeed Insights tool. With that came a fresh new design and the following release notes:

“v5 of the PageSpeed Insights API was released in November 2018. It now uses Lighthouse as its analysis engine and also incorporates field data provided by the Chrome User Experience Report (CrUX). v5 of the API will now provide CrUX data, and all of the Lighthouse audits.”

It’s not just a new look; under the hood, the new PI now incorporates, as mentioned in the release notes, lab data by Lighthouse and field data gathered from Chrome User Experience Reports – so what does that mean exactly?

Essentially that the tool is no longer simply a scan of a website’s code and a set of recommendations based off that, but instead the analysis is now also backed by reports on a site’s real-world performance.

Lighthouse is a familiar web speed audit in-built with Chrome while CrUX metrics detail the experience that real users have had with a page over time when using the Chrome browser.

Google PageSpeed Insights – the old vs. the new

PageSpeed Insights, for those still unaware, tells us how a given page is performing speed-wise on both mobile and desktop devices and provides suggestions on how that page may be improved. Along with a report will be a performance score – 90 or above is considered fast, 50 to 90 average and below 50 (worryingly) slow. That remains unchanged.

In terms of usability, it’s also still the same easy process of entering a website’s URL in the top bar and letting the tool go to work.

How it looked previously:

Compare that to the below, a snapshot of the results page now:

What’s changed in the results?

You’re given results tailored to both desktop and mobile in separate tabs, just like before.

The results page, however, delivers its recommendations in a few new sections – ‘Opportunities’, ‘Diagnostics’ and ‘Audits’. Within those are several of the same checks carried over from the old tool including the likes of browser caching, redirects, compression of images and minification of JavaScript/CSS/HTML. After all, they’re still important.

On top of those are some all-new optimisation tips.

What might be the biggest difference between the old tool and the new is that it has been reported that even if your site has the sought-after score of 100 for page speed, you may still see a list of opportunities to improve. No more resting on your laurels when it comes to page speed.

But does speed really matter?

In a word, yes. That’s why we’re here!

Speed is now a ranking factor, and that makes improving it a crucial component of SEO. Google’s July 2018 Speed Update algorithm change saw to that.

The importance of speed is not only apparent on desktop sites but mobile too.

And with the new PageSpeed Insights, those issues really can’t be avoided, as we’ll see.

You may also like – 10 Google Search Operators to Mine Search Results

Small load times equal big business this Christmas

As discussed in our last insight into pagespeed Amazon once calculated that a page load slowdown of just a second could cost the company $1.6 billion in sales each year – hopefully, they’ve got that in check for the festive period!

Sorting out your speed can start with a few simple changes, such as:

• Improving your server response time
• Combining any external JavaScript and CSS files
• Enabling gzip compression
• Building AMP pages (more on this later)

But don’t take our word for it straight from the horse’s mouth.

Putting Google PageSpeed Insights to the test

Time to put the newly-designed PageSpeed Insights into action, side-by-side with the previous one.

Well, Google is making the old version a little tricky to find as naturally they now direct users to the shiny new one (and they’ve stated that v4 of the PSI API will be deprecated within six months).

Time then for a sneaky workaround.

If you run a scan of your site on testmysite.withgoogle and then have it email you a more in-depth report that will, at the time of writing, send you a list of recommendations based on v4 of Google’s PageSpeed Insights.

Okay, it’s not identical in design to the original tool but it covers the same speed criteria, and each point is linked to the corresponding v4 documentation.

Choosing the test subjects

American Airlines

Aeroplanes are fast. So let’s test americanairlines.co.uk.

‘Test My Site’, by the way, looks specifically at mobile speed so we’ll compare it to the results in the mobile tab of the new PI (and hint: we hear mobile is what has changed the most in the tool).

americanairlines.co.uk with v4 API recommendations via ‘Test My Site’:

americanairlines.co.uk tested on the new version:

Although no score is available now on the older tool using the v4 API, we can observe that the old and the new both report americanairlines.co.uk to have poor/slow page speed on mobile.

Look at all that red.

The site passes several audits such as server response time, offscreen images and minification of JavaScript but falls down on tests for render-blocking resources, redirects and minification of CSS among others – with the top opportunity potentially slashing speed by almost 5 seconds.

Tyson Gay

We’ll also test the website of Tyson Gay – the second fastest man in the world – because hey, Usain Bolt should have a speedy website already.

tysongay.com tested with the v4 API:

tysongay.com tested on the new version:

More snail’s pace speed reported for mobile.

With speed now a part of Google’s search algorithm, technically a speedier competing website could outrank yours. Mobile searches have surpassed desktop so speed matters on these devices.

A slow speed like this may affect visitor retention – today’s users are demanding and could abandon a site if it takes too long to load. A slow site is not an inviting one. Cut the load time by just a fraction of a second and it could increase the likelihood a user will stick around and explore some more.

Comparing the results

The complete version of the previous PageSpeed Insights would have given us more a bit more detail than what we had on hand here for the testing, but the same core list of recommendations.

How do the findings differ between v4 and v5?

To put it simply v5 covers everything v4 did and more. All of the recommendations of the old tool in the two tests were carried over to the new, including the likes of:

• Image compression
• Resource minification
• Browser caching
• Eliminating render-blocking JS/CSS

Those are now joined by further recommendations not present before, such as:

• Ensuring text remains visible during web font loads
• Avoiding excessive DOM size
• Minimising main-thread work
• Minimising critical requests depth

All sounding a bit technical, eh? Luckily, there are help pages like this linked with the recommendations. And although the pointers in the test results contain a lot of tech-speak they are backed with some impressive data to look at by way of examples found in your website.

The pointers feel more up-to-date and forward-thinking than before too:

Not forgetting on top of this the ‘Lab Data’ delves deeper than solely the First Contentful Paint measured in the previous tool, and ‘Field Data’ brings in real-world results gathered from Chrome users which wasn’t there before.

But you’ll need to have a pretty popular site to see that ‘Field Data’ section; it worked for americanairlines.co.uk, but not for tysongay.com.

Which is better?

Predominantly the old and new versions work with the same core list of checks and recommendations. However, v5 expands upon them with new pointers that feel more relevant to the times, even if a little too technical for the average webmaster to decipher.

Of course, the major upgrade here is the introduction of lab data that would prove useful for debugging performance issues and field data that helps capture true user experience.

Mobile affected the most

Feedback on the new tool suggests that mobile speed is now scored more harshly than before across the board while on desktop sites it’s easier than before to score a perfect 100 – it certainly seems that way going off subsequent tests we made on other websites.

According to discussions on Centmin Mod v5 is heavily weighted towards Time to Interactive, Speed Index and First Contentful Paint (metrics in the ‘Lab Data’ section), resulting in lower scores for dynamic and JavaScript-heavy sites in the 3G emulation mobile testing.

The implications

Is the more stringent testing on mobile leading to the inevitable? AMP.

Accelerated Mobile Pages (AMPs) deliver mobile pages fast, and Google has been a big endorser of them.

These stripped-down versions of your site pages remove anything unnecessary that would usually slow them down. Google pre-renders and loads AMP pages instantly. And Google likes that.

AMP pages fare better in testing tools like this (…you could say v5 was built with AMP in mind) and they may well give you a boost in rankings.

The need for speed

It’s more important than ever to tackle the speed of your website, and as we’ve seen, that’s particularly true for mobile.

If it can help you rank over a competitor, which it can, then it’s crucial to your SEO efforts.

Check your speed and start planning for the new year. Google PageSpeed Insights is a great place to start – but don’t forget GT Metrix and Pingdom because it’s important to get as many opinions as possible since figures can fluctuate between the different test sources.

You’ll also be amazed what you can measure and debug using Chrome Dev Tools built directly into Google’s browser. Hit F12 and explore.

The need for (improved) speed is unavoidable, and it applies to everyone, even Usain Bolt!

Get in touch to learn about our complete Pagespeed Audit service and give your website that much-needed boost!