tag:blogger.com,1999:blog-537292020947772033.post6036388959300083083..comments2023-05-26T04:28:47.893-04:00Comments on Performance Matters: Measuring performance of the user experiencePatrick Meenanhttp://www.blogger.com/profile/16482126817753317557noreply@blogger.comBlogger8125tag:blogger.com,1999:blog-537292020947772033.post-15917401849806103292014-02-17T12:43:26.525-05:002014-02-17T12:43:26.525-05:00And there you have one of the reasons :-). I'...And there you have one of the reasons :-). I've seen cases where the clocks went backwards so you'd get negative numbers (shouldn't happen since it's supposed to be a monatonic clock) and also cases where sometimes a browser version would use the wrong time values for different times (I know Chrome had a bug for a release or two where it used time() instead of performance.now() which were radically different - one being a timestamp and one being relative to nav start).<br /><br />There are also some edge cases on mobile where a browser goes to the background during nav, gets paused, comes back at some point in the future and ends up with a crazy-long time.<br /><br />Dropping the crazy outliers works better than capping them to a max value because they will still skew averages pretty significantly.Patrick Meenanhttps://www.blogger.com/profile/16482126817753317557noreply@blogger.comtag:blogger.com,1999:blog-537292020947772033.post-60575168886159556882014-02-13T19:25:11.813-05:002014-02-13T19:25:11.813-05:00Great stuff Pat. I was curious why it is necessary...Great stuff Pat. I was curious why it is necessary to filter out numbers outside of these ranges:<br /><br /> if (t >= 0 && t < 3600000) {<br /><br />I didn't filter them out, and IE9 showed quite a few values that looked like timestamps. <br /><br />Jamishttps://www.blogger.com/profile/03054474807182384219noreply@blogger.comtag:blogger.com,1999:blog-537292020947772033.post-35391098543783422892013-10-04T14:17:55.300-04:002013-10-04T14:17:55.300-04:00As confusing as it is, we've been calling it a...As confusing as it is, we've been calling it aft for "above fold time" for years. I should change the part in the parenthesis - thx.Patrick Meenanhttps://www.blogger.com/profile/16482126817753317557noreply@blogger.comtag:blogger.com,1999:blog-537292020947772033.post-60806019271373861282013-10-04T13:37:54.356-04:002013-10-04T13:37:54.356-04:00Great article and will share. Just a note for auth...Great article and will share. Just a note for author, I believe there is a typo.<br />"aft (above-the-fold)" should be "atf (above-the-fold)"<br /><br />Right? I would want someone pointing that out to me. All else, great article. Thank you for writing it.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-537292020947772033.post-8790932382898643202013-08-01T10:55:57.496-04:002013-08-01T10:55:57.496-04:00Good stuff!
We've implemented our own beaconin...Good stuff!<br />We've implemented our own beaconing of events straight into graphite for easy graphing: http://spil.com/makingrumactionable Frank van Gemerenhttp://www.spilgames.comnoreply@blogger.comtag:blogger.com,1999:blog-537292020947772033.post-89663206441291435012013-08-01T08:39:30.907-04:002013-08-01T08:39:30.907-04:00You have some great info here. However, the black ...You have some great info here. However, the black background color makes it practically impossible to read. Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-537292020947772033.post-14287912780062853152013-07-15T17:02:22.568-04:002013-07-15T17:02:22.568-04:00Awesome post. Getting custom user timings into fre...Awesome post. Getting custom user timings into free tools for RUM like GA is a huge win for many organizations that struggle with the expense of ongoing performance testing and analysis (non-profits I'm looking at you). Thanks for making sophisticated measurement more easily available to those who need to leverage tech to the hilt the most. Jennifer Showehttps://www.blogger.com/profile/06187134133122752468noreply@blogger.comtag:blogger.com,1999:blog-537292020947772033.post-65891647941739701782013-07-15T16:35:18.940-04:002013-07-15T16:35:18.940-04:00Timers that reflect what the page is trying to do ...Timers that reflect what the page is trying to do and how users perceive that is critical. Thanks, Pat, for leading the charge. The shim to send User Timings to GA, mPulse and Boomerang is SWEET!Steve Soudershttp://stevesouders.com/noreply@blogger.com