This is real world page speed optimization

  • 7 October 2021
  • 4 replies
  • 110 views

Userlevel 6
Badge +3

This is a real world optimization!  This is an example of a site which, I confess, I’ve struggled to get into the green on GTM and PageSpeed Insights lab results.

Now, finally, the site has crossed whatever bridge Google has put in place and started showing Field Results.  As you will see, below, the site passes field results the core web vitals assessment with flying colours for both desktop and mobile but, lab results still look pretty dismal but, they are now meaningless!

Does anyone know what traffic needs to be present before Google shows field data?

 


4 replies

Userlevel 3

To your original question… I have no exact number on how much traffic is needed for field data to be presented / considered by Google but as a point of reference I have a website which gets around 2000 hits a month (so not a huge amount) and some sites which have even more than that (upwards of 8,000) and they all show the field data from Google, but then I have a site with about 1400 hits a month which has no field data, so I am making a wild guess as it’s maybe somewhere around the 2000 hits/month mark?

 

Rant ahead... The PageSpeed Labs or Google’s web.dev site are horrible IMO as they simulate an incredibly slow mobile speed (slow IMO for most developed countries anyways). I often just use Lightspeed right in Chrome as that’s far more accurate in my experience (and webpagetest.org too). For example, I may get only a score of 81 on web.dev / PSI but in Lighthouse via Chrome I’ll instead get a 94. I recognize many parts of the world still have slow mobile speeds, but many developed countries will have at least 15 Mbps to upwards of a few hundred Mbps (either at home or on LTE).

I could be a bit biased of course because the vast majority of the websites I manage and built are meant for local audiences (i.e. small businesses that only ship locally or in the country of business, for example) and our speeds here in Canada are pretty great for mobile LTE, which is why the PSI or web.dev scores are so inaccurate to me because they’re simulating something that simply isn’t realistic to the local audience where my customers operate in. I suppose web.dev may be more realistic if a website truly has an intended world-wide audience. :shrug:   Anyways, felt I needed to rant a bit on that. haha. /rant

Userlevel 2

I have one site 2.5k visits a month - no field data - I have one with 2k - field data - so I have no idea what the trigger is

 

Userlevel 3

Keep in mind this field data is only taken from Chrome

users. So if the majority of your web visitors are using Safari for example, then you likely won’t get enough for field data. 

Userlevel 2

Ahhhhhh that is an interesting point - and almost certainly explains the difference between the two sites

Reply