As a website owner, you know your site needs to be fast and now you want to know how your site is doing. So you head over to Google PageSpeed Insights, because that’s what all the articles and tutorials on the web tell you to do, and you enter your site's URL. You’ll be presented with a grade and a list of recommendations from Google and at that point, you might be dismayed:
- “What are all these red and orange warnings?!”
- “Why isn’t my grade higher?!”
- “What do all these recommendations mean?!”
You might be expecting that your PageSpeed grade will be near-perfect. Or you’ll look at the recommendations and wonder why your developer hasn’t fixed them all.
Clients sometimes ask why their site’s PageSpeed grade isn’t higher, or they assume that because it didn’t increase a lot, it must mean their developer doesn’t know how to make a fast website.
The simple truth is this:
Your Google PageSpeed score does not matter.
The Need for Speed
Speed, i.e the loading time of your site is the most important metric. This is what counts for user experience and for SEO. When the Google bot crawls your site, it cannot see your “grade”, only your speed.
Did you know that Google PageSpeed doesn’t even measure the loading time of your site?
Think back to your school days. Did perfect grades mean you were smart? Not necessarily. It just meant that you knew how to do well on tests. But many intelligent people simply not do well on tests.
So just like school grades are not an indicator of intelligence, Google’s PageSpeed grade is not actually an indicator of speed.
Google PageSpeed does not actually measure the speed of your site.
Here are 3 websites all with similar load times, but with vastly differing PageSpeed scores:
|Site||Load time||Pagespeed Score|
For each of these 3 sites, the loading time is under 1 second (which is fantastic) but the PageSpeed scores range from 58 – 91.
And the below site has a really good PageSpeed score yet is significantly slower than all 3 of the sites above:
|Site||Load time||Pagespeed Score|
So you can see from these examples that the Google PageSpeed grade is not an indicator of speed.
Even Google doesn't score well!
If you’re still not convinced then this spreadsheet should seal the deal. We personally tested 46 of Google’s own websites and recorded their PageSpeed scores with their own tool. They didn’t fare well.See for yourself!
Chasing a grade is a waste of time
No site gets a perfect grade, in fact it’s pretty much impossible to achieve, and since it doesn’t correlate to speed, why bother?
If you try an attain a perfect grade, by implementing all the suggestions Google PageSpeed makes, you will lose your sanity pretty quickly.
You cannot take too literally all of the suggestions from Google PageSpeed because sometimes they are unrealistic or impossible. For example, it may tell you to minify or add expiry headers to a file that is not hosted on your website and beyond your control.
In the below example, only Facebook and Google themselves can add browser caching to these files.
If a CDN is used for your site, PageSpeed might give you a lower score for that, but in many cases a CDN provides better speed for your visitors. Or if it is marking you down because you could save 1kb by compressing an image (see below), it’s not worth bothering about!
It typically then provides 2 parts to this recommendation:
Optimize CSS delivery of the following
Google suggests you can “Optimize CSS delivery“.
Essentially they want you to separate out the CSS needed to render the first part of your page and embed it directly into the code of your page, rather than in the main stylesheet.
So as you can gather, if you are not a developer, this is pretty hard to do and involves re-coding parts of your theme. It will also vary from site to site, so it’s not something that can be made into a one-click solution in a Wordpress plugin.
A developer can implement this technique. It won’t necessarily improve the overall load time of your page, but it will increase your PageSpeed score, and may appear to load faster because the elements at the top of the page will load sooner. This type of micro-optimization is utilized by sites like Amazon where a 10ms difference may impact their bottom line. But for most small sites, it’s unlikely to have much impact.
So what is PageSpeed good for?
Google PageSpeed can be helpful as long as you don’t treat it as the be-all, end-all.
Sometimes it can alert you to problem areas on your site that can be addressed. For example, it might alert you to the fact that your content is not being GZIP-ed.
Or it might alert you that you have too many large images which could be compressed. This is a good recommendation which we can actually act on.
Below is a great example of an opportunity to improve because you can save 750kb, an 87% reduction in size. This would have a good effect on your site’s loading time:
So it’s best to look at PageSpeed as one of several tools that might provide some pointers, but your goal should always be to improve actual speed, not your “PageSpeed” grade.
Guidelines for using PageSpeed
- Do not blindly trust Google PageSpeed or take it at face value.
- Always read the recommendations carefully and assess if they are possible and worth the time and money. If it’s asking for something impossible, you should ignore that!
- Don’t forget to always focus on speed and don’t worry about chasing a grade.
- Always use an actual speed testing tool like Pingdom to see the impact of any changes made on your site.
Skunkworks always aims for maximum speed for our client’s websites and we are capable of building websites for clients that load in under 1 second. If you want to improve your website’s speed talk to your account manager to find out more about what would be involved with getting you to where you want to be.