Did something similar for my SmashingConf talk - basically grab timeline JSON for HTTPArchive run, and run it through 3rd-Party-CPU abuser (https://github.com/andydavies/3rd-party-cpu-abuser …)
@paulcalvano’s approach might be quicker though
-
-
Replying to @AndyDavies @addyosmani and
@AndyDavies 3rd-party-cpu-abuser modification is super useful. I occasionally combine it with an analysis of requests and byte sizes from WPT. For example, how much 3rd party content loaded before start render, or first meaningful paint, or in this example before a interstitial.pic.twitter.com/oMQhhNPytC
1 reply 0 retweets 5 likes -
Replying to @paulcalvano @AndyDavies and
There are a few
@HTTPArchive discussion forum posts on 3rd party content - and we can do some deeper analysis to try to understand how performance trends based on % of 3rd party content. Would be really interesting to combine HA analysis w/ CrUX response times for this...1 reply 0 retweets 3 likes -
Replying to @paulcalvano @AndyDavies and
This graph illustrates the ranges of load times for the top 10K Alexa sites. I used
@igrigorik's regex for 3rd party classification (https://discuss.httparchive.org/t/what-is-the-distribution-of-1st-party-vs-3rd-party-resources/100/11 …) + CrUX onLoad metrics. There is definitely a strong correlation between % of 3rd party content and load times!pic.twitter.com/MSmDieqdKE
2 replies 4 retweets 18 likes -
Replying to @paulcalvano @AndyDavies and
So great to see this with CrUX data factored in. Thanks for investigating, Paul!
1 reply 0 retweets 3 likes -
Replying to @addyosmani @paulcalvano and
What metric was being used for "load time" here?
2 replies 0 retweets 1 like -
Replying to @slightlylate @paulcalvano and
CrUX has FP, FCP, DCL and onLoad (no TTCI/FID just yet). We can cross reference origins with their metrics in Lighthouse from WPT (TTCI) which should show sites with higher TTCI are referencing more third parties.
2 replies 0 retweets 7 likes -
Replying to @addyosmani @paulcalvano and
yeah, I expect the correlation to hold (strongly) as we use more-inclusive metrics (TTI/FID). Was just wondering how so many of those reports seemed to show < 2s ;-)
2 replies 0 retweets 4 likes -
Replying to @slightlylate
What’s FID? guessing First Interactive something?
1 reply 0 retweets 0 likes -
Replying to @AndyDavies
First Input Delay:https://developers.google.com/web/updates/2018/05/first-input-delay …
2 replies 1 retweet 3 likes
Basically, we were never able to agree about how to measure TTI well in the field, so FID is our closest proxy metric.
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.
& Web Standards TL; Blink API OWNER
Named PWAs w/
DMs open. Tweets my own; press@google.com for official comms.