When you are shopping for a new vehicle, most manufacturers list the 0 to 60 mph times for their vehicles. This metric serves as a proxy for the build quality, engineering expertise, power, handling, weight, and aerodynamics of a car. You can also compare two completely different classes of vehicles based on this metric and still get usable information.
For example, imagine having these cars on your shortlist:
Subaru Outback (0-60 in 5.7 seconds)
Honda Accord (0-60 in 5.6 seconds)
VW Passat (0-60 in 10.7 seconds)
Which car would you remove from the shortlist first if speed was essential to you? Probably the Passat, right?
Now imagine if all cars listed their metrics like this instead:
Subaru Outback (Handling: A, Power: A, Safety: B)
Honda Accord (Handling: A, Power: A+, Safety: A)
VW Passat (Handling: A+, Power: A-, Safety: A)
Would you pay attention to those grades? I guess not.
Yet, when it comes to WordPress and websites in general, we insist on using letter grades to measure the speed. Most developers use either Google Pagespeed Insights, Pingdom, or GTMetrix speed test to measure site speed.
All 3 of these testing services have chosen odd metrics to highlight prominently on their reports, which causes the community to focus on wrong metrics.
Pagespeed insights report shows site speed on a scale of 0 to 100. Pingdom assigns a performance grade from A to F. GTMetrix assigns two letter grades to the site, one for pagespeed and one for yslow.
This leads to two types of posts on WordPress forums and Facebook groups. One where developers are worried about going from a C to A for a site that loads in 3 seconds. And second, where developers showcase their perfect pagespeed score of 100 when the site, in reality, takes 5 seconds to load.
While all 3 of these testing services display the site speed measured in seconds, it’s listed as a second class metric that doesn’t seem important. Apart from the odd metrics, these results can be manipulated quite easily just by adjusting testing locations and throttling.
So how do you measure WordPress site speed accurately?
If you are running a high traffic WordPress site, you can’t go wrong with real user monitoring tools like New Relic and Dynatrace. While these tools are a bit expensive, they provide in-depth visibility into every visitor’s experience. The ability to continuously monitor site speed across browsers, devices, and regions empowers your WebOps and dev teams to keep the site running fast proactively.
If using RUM tools is not an option, you can get usable insights from tools you already use, such as Google Analytics and Facebook Pixel. Change the sample size to 100% in Analytics, and you can see the site speed metrics across user sessions, browsers, regions, and even specific pages.
So are the site speed tools completely useless? Not if you focus on the right metrics and have a solution in place to measure real-world performance as well.