Google PageSpeed Vs. Google Lighthouse

Result-driven website development is an absolute for me as a freelance web designer/developer. If my project gets results, I get a good review and most importantly I get more work.

Getting tons of top ranking pages in Google organic search is an awesome feat. This win is typically derived from a combination of consultation, project planning, and development. One shortcut to a high-ranking site is high Google scores through Goole's website auditing tools.

Throughout the past few years, Google Lighthouse has been my go-to for auditing my website for performance results. Getting the nod of approval straight from Google is gold. They are also the toughest critics. Lighthouse has been available to anyone through the Chrome browser; however, it's pretty hidden.

Recently, Lighthouse has rolled out to the general public in a secondary form called Google PageSpeed. Thing is, its fickle. Let's go over how and the potentially justified reasons why.

Case in point: 3200.Pro Stats

test

Google Lighthouse: In Google Lighthouse I get 100% on both Mobile and Desktop. Pretty awesome place to start a website. I also get to watch some awesome animated fireworks. (Side note, I didn't start seeing these until I started working with the JamStack)

Google PageSpeed: I ran the test 6 times in private browser windows. Once Google gave me a 'B', shortly after an 'A+'. Like I said. Fickle. But, after reading through the docs at Google there is a method to the madness and a suggestion on how to work with it.

Justification: The reason for the variance is a simulation that they run to give the sites scores variability of performance.

According to Google 'Variability in performance measurement is introduced via a number of channels with different levels of impact. Several common sources of metric variability are local network availability, client hardware availability, and client resource contention.' Personally I'd prefer they don't because I have to explain this perpetually but It makes sense.

What to do about it: Here's a quick list of things you can do to make it a little easier when working towards your maximum optimization on client projects:

  1. Use the Chrome inspector version of Lighthouse Audit. It uses Google Lab standards that don't vary from what I can gather.
  2. Explain to clients that they should run pagespeed 5-10X to get a good range of scores. The variance can swing 20%+ and the highest score you can get is legit.
  3. According to John Mu (Webmaster Trends Analyst at Google) relax about the final grade. Keep an eye on the itemized report of what you can improve beneath the grade.

...

Test

© 2020 3200.Pro | Site Design: 3200.pro