Page Speed Ponderings

Just some pondering about web pages speeds. What really matters and what doesn’t?

This is not meant as a tutorial and I’m not an SEO expert. I’m just commenting on what I see and my interpretations.

I build and manage websites for small businesses in Australia, that have clients in Australia. My end goal is to make sure pages load fast in Australia. So hosting here just makes sense. However, a lot of tools and analysers test from other countries and do not take into account the internet latency between them and us. In our case often 600ms to 1000ms. So some clients get reports from marketers showing them poor scores and page speeds from Google PSI, which is completely wrong.

Google real-world Core Web Vitals takes quite a bit of traffic before any metrics are made available. So people often refer to the emulated Page Speed Insights scores. PSI does give some good insights into possible issues. However, the actual scores in my view are ambiguous and affected by other factors, like the location of your hosting etc. Google also state PSI is not a ranking factor.

There are also ways to fool Google PSI and other emulators. Some Caching tools serve up completely different pages to PSI, GTMetrix, Pingdom etc, than to actual users. The pages served to the tools have no JS, Limited CSS and reduced image sizes. So they achieve a great score but are not really getting scored on what the user’s actual experience is. Great for marketers selling SEO, by what real difference does it make?

BTW: Got to love demos. My internet speed dropped while creating this video. Page load times in Chrome was over 2 sec, prior to the video they were around 0.8sec. However, you can see in the GTMetrics tests from Sydney, the FCP is under 0.8s;