Chasing a 100/100 score in Google PageSpeed Insights (PSI) is a waste of time. Today, I will explain why PSI scores don’t really matter and why you shouldn’t completely trust site speed testing tools.
First of all, don’t get me wrong, website speed matters. You should still optimize your pages to offer fast loading times and a positive user experience.
Here are 5 reasons why you shouldn’t be obsessed by PSI scores:
1. Core Web Vitals is just a reference
The current performance assessment is based on 3 aspects: loading, interactivity, and visual stability. These are represented by the following metrics:
- LCP – Largest Contentful Paint, should be 2.5 sec or less
- FID – First Input Delay, should be 100 ms or less
- CLS – Cumulative Layout Shift, should be 0.1 or less
While these metrics are important, the real-world user experience beats them all. Your main goal is to offer blazing fast web pages, an appealing UI (user interface) and a positive UX (user experience).
Get feedback from your real visitors or customers and try to deliver what works best for them. I’ve seen websites with a perfect 100 score, but mediocre design and content.
2. The tests are based on a simulated environment
For mobile scores PSI simulates the test for a mid-tier Moto G4 device and using a 1,638 Kbps connection. Most of your users won’t be visiting your site from an old device and a slow Internet connection.
In fact, according to Speedtest, the average global Internet speed on mobile devices is 28.61 Mbps for download and 8.38 Mbps for upload.
3. The Scoring Calculator is changing frequently
The current version of Lighthouse Scoring Calculator is 8.0.0. It calculates the score based on 6 metrics. Every metrics has its own weighting. This weighting is changing with every new version of the calculator.
Let’s say you worked hard to improve your FCP (First Contentful Paint). In version 7 its weighting was 15%. In version 8 its weighting is 10%. According to this calculator your performance got worse.
Don’t try to please Google’s speed testing tool, try to please your real users.
4. Ignore lab data, focus on field data
Google PSI measures the performance in 2 ways: in the lab and in the field.
Lab data is based on simulated tests. These are the scores you get in the testing tool. These are vanity scores, simply ignore them.
Field data is based on data from actual users around the world. This data is more relevant. According to field data you either pass or fail the Core Web Vitals Assessment.
The problem is: you need enough traffic to get an overview of your site’s performance based on real visits. According to my estimations, you would need at least 60-100 visits/day or about 2-3k visits/month.
Most small businesses won’t have this amount of traffic => no data on performance.
5. Mobile usability report is limited
Even if you have a website with at least 2-3k visits/month, you still need to have fully valid mobile usable pages. The issue is that Google won’t evaluate all your web pages, but only a limited number.
So if you have a large website, Google will only mark a small number of pages as fully valid (up to 50% based on my experience) in the Mobile Usability report. This means that your performance is assessed based only on those pages.
This is again, not accurate and doesn’t represent the full picture on performance.
That’s it friends, my conclusion is: don’t chase the perfect score. Focus on delivering the best user experience to your visitors/customers, based on their needs.
What do you think – do Google PageSpeed Insights scores really matter? Let me know in the comments below.