Good Morning, Yannick Gaultier,
once again thank you for your help to make 4SEO running. Now I have some first data - and another question.
How the Google-Speed Data are calculated?
Expacted behavior would be the average or the 75th percentile. So I'm a little bit puzzled about some warnings. For example one simple page.
https://www.example.de/trauertexte/zitat-spruch/seemannsgrab-schmuck-moewen
4Seo reported Problems in all 4 Areas:
It states FID: 260.300, INP 304.000, LCP 2.705.699, CLS 0.5903
Base for it are 3 measures. In the phpMyAdmin I found them with the following values
FID: 1.400, 260.300 and 31.000
INP 24.000, 304.000, 136.000
LCP 1.530.099, 2.705.699, 855.500
CLS 0.025234, 0.5903 and 0
So its obviosly not the average of 3 measerements shown, but only the peak. In this case all peaks are from the same visitor, in an other case it is the peak from various visitors (url trauerkarten-bank-baum)
The average would be inprove CLS, the rest is fine.
So the reported erros are not the expacted behavior (the average or the 75th percentile), but the maximum, created by one slow connection. Enclosed you'll find the tables extracted from my SQL.
Another helpful filter would be to separate mobile from desktop-visitors as Google does. You have the data in perf_data, so I can use them via SQL, but would be more convenient in the app.
Thanks a lot anyway for your great App. It already helped me fasten up my page.
Have a nice day and best regards from Germany
Jürgen Kahle
Helpdesk is open from Monday through Friday CET
#10956 – Question about Google Speed LCP FID INP CLS
Hi
Expacted behavior would be the average or the 75th percentile.
That's not how CWV are computed. They are not the "average of the 75th percentile". They "are the 75th percentile".
So to pass a CWV metric, you need to have at least 75% of your visitors with a value above the minimum, or the opposite: you cannot have more than 25% of visitors below the threshold.
So with 3 measurements, you need at least 3 * 0.75 = 2.25 visitors that are "good", but in your example you fail because you have only 2 out of 3.
Obviously, that does not work very well with only a few measurements. However, keep in mind:
- this is what Google uses. Exactly.
- they may or may not have more measurements than 4SEO
Best regards
Yannick Gaultier
weeblr.com / @weeblr
Thanks for the quick respond.
You're right. I checked it with some more data - thank you.
So my only wish is to seperate mobile/desktop in a future version and I'm happy :-)
Thank you very much and have a nice day
Jürgen
Hi
So my only wish is to seperate mobile/desktop in a future version and I'm happy :-)
The reason it's not split simply that it would require more space. Or a separate report entirely, which is some undertaking.
I have a tracker item to improve the CWV reporting, it's just not gotten to the top of the list so far.
I don't want to do a full dedicated report here, as the results we measure are "just" a warning and a detector for pages you may have to work on. The values by themselves are not that important.
The goal here is that 4SEO alerts you of a problem on a page. You're not going to use the 4SEO values to actually work on the page. You're going to use Google Page SPeed insights or your browser dev tools to try and understand what's causing an issue - and for which users - and then hopefully fix it.
That's why I did not considered it absolute required to show Overall/desktop/mobile alerts.
But it can certainly be done to show the details on the page details for instance.
Best regards
Yannick Gaultier
weeblr.com / @weeblr
