Google PageSpeed, GTMetrix, YSlow Speed scores … how important are they?

People think that these metrics are an ironclad predictor of how fast and "good" your site is.  They aren't.

First off, these metrics have nothing to do with how fast the page loads for the user.

Yes that’s right … the page might be loading instantly for the user but the metrics might be looking for specific things that it ticks off as essential for meeting it’s “metric” for loading.   Conversely the site might be painfully slow to load, and provide a terrible user experience, but it might rank very highly in Google PageSpeed or some of the other tools. Think about this for a second. You are dealing with intelligent (some of them at least) emotional human visitors to your website. You are not dealing with robots who can’t see or hear (yet).  There have been some strides in AI but we are nowhere near having an AI with feelings, perceptions and emotions that you would need to rank how fast a site is perceived to load.  Albert Einstein once famously replied when asked to explain relativity to a reporter :

“Put your hand on a hot stove for a minute, and it seems like an hour. Sit with a pretty girl for an hour, and it seems like a minute. That’s relativity.”

The point is that the totality of the site might be slow to load but the user won’t perceive it as such because the content they are focused on loads fast. Also the AI robots Google and the others use can see things humans can’t see, and can rank things that in a lot of cases make no difference to visitors perceptions of speed.

People think that these metrics are an ironclad predictor of how fast and “good” your site is.  They aren’t. In fact the main value they have is that your Google PageSpeed score affects the ranking of your site with Google.  Other than that, there are sometimes some useful tips you can get from them, but at the end of the day, your own experience is the best metric to judge the page loading of the site, not some automated tool. If you load a page and it loads quickly and looks good, this is the best measure of how good your page speed is.  We all know those people who are brilliant at exams but hopeless in the real world. This is a similar phenomenon.

If you run the PageSpeed tool on some of the largest sites out there with vast resources behind them, you will get surprising results. Below is a table of the scores of the first 3 sites that popped into my head. Go ahead, try some more, and let me know if you find faster ones:

Site PageSpeed Score 1 14 53

Very few people would argue that these sites are “slow” or deliver a poor user experience by any subjective measure. And the subjective measure IS THE ONLY THING THAT MATTERS to the person viewing the page.  Visitors expect a rich experience when they visit your product or service site.  The site that will score the best on the speed metrics would be one with “Our product is great, buy it now” in black in Times New Roman on an otherwise blank page.

Improving the score on one often means decreasing the score on another.

For an optimisation on a particular site, you will often see this happen. So I would suggest focusing on Google PageSpeed.  I would also not worry about getting 90%+ scores, as the incremental cost of this is exorbitant, and the effect on the user experience is terrible. I would aim for anything above 30%. Since Google themselves say:

“The “Speed Update,” as we’re calling it, will only affect pages that deliver the slowest experience to users and will only affect a small percentage of queries. It applies the same standard to all pages, regardless of the technology used to build the page. The intent of the search query is still a very strong signal, so a slow page may still rank highly if it has great, relevant content.”

Google’s Zhiheng Wang and Doantam Phan

One of the main things recommended for speeding up your site is “Lazy Loading” of images. The only problem is that “Lazy Loading” images makes most sites look terrible.  If you have a lot of images in your site – as most sites these days do – then you end up with loads of blanks with the odd line of text here and there, and then the site “jumps around” like crazy when the images load. This looks awful.I am sure many of you have seen this happen.  There are ways around it but it takes custom work for each site.

In conclusion, I would say that the best strategy is to ensure that your site is perceived as fast to visitors, but not to get too hung up on your PageSpeed score. We can certainly improve it but we won’t be aiming to make it 100%.