Site: US UK AU |
Nexcess Blog

Pursuing A Perfect PageSpeed Insights Score Isn’t The Best Use Of Your Time

May 4, 2016 0 Comments RSS Feed

PageSpeedPageSpeed Insights is a useful tool from Google that gives website owners a “score” that represents how well optimized for performance their site is. It also displays a rundown of the ways in which a page’s performance optimization falls short: render-blocking JavaScript or improper use of browser caching, for instance. The score is generated by adding or subtracting points depending on whether specific optimization best practices have been implemented on a site. As an educational tool, it’s invaluable, giving non-expert users a list of what — in theory — they need to do to get their site in tip-top condition. But, in spite of its usefulness, I often find myself warning users not to take their PageSpeed score too seriously.

The reason is simple: PageSpeed Insights does not measure a site’s loading speed, it simply categorizes it according to a series of performance best practices and produces a score. The problem with this approach is that performance optimization is a balancing act that combines many different factors. No single score can encompass the totality, and so that score can be misleading.

In fact, it’s entirely possible for a site to have an excellent PageSpeed score, and yet load more slowly than a site with a lower score, as was demonstrated nicely by Lucy Beer on the WP Rocket blog. Following the advice of PageSpeed Insights is broadly a good idea, but the advice is general and the factors are weighted such that a trivial infraction that does little to impact real world performance on specific sites can have an outsized effect on the score.

The second problem I have with PageSpeed Insights is that often the advice it gives simply isn’t practical. Let’s look at an example: if a webpage loads a render-blocking JavaScript file in its head section, it will never achieve a perfect score. This is, in general, sound because render-blocking scripts can seriously degrade a site’s load times, but sometimes you need the functionality of that script on the initial page load. There are ways around it, but since PageSpeed doesn’t tell you whether the render-blocking script is actually causing a meaningful delay in page loading time, you can’t determine whether it’s worth the bother of fixing.

The problem is compounded if you happen to be non-technical user who just bought a WordPress theme and wants to know how well it performs. Firstly, the advice Google gives is not advice you’re likely to be able to implement without coding skills, and secondly, you don’t know whether it’s worth implementing in the first place.

So what’s a site owner to do? I want to make it clear that I think PageSpeed Insights is an awesome tool, but only if you take the information it gives in context and understand that the aim is better performance not a better PageSpeed Insights score — they aren’t the same thing.

Site owners should complement PageSpeed Insights with a tool like Pingdom Tools, which will also give you a score, but, more helpfully, tell you how long a page took to load from various locations around the world and a graphical representation of which elements are taking the most time to load.

Posted in: General