This task is to revisit reports from MediaWiki policy on IE11 that were made in 2021 around who is using IE11, what behaviors they engage in, and what countries they are from. With IE11 reaching end of life on June 15, 2022, we should revisit its prevalence among WMF users and consider moving IE11 to unknown support. S
Description
Status | Subtype | Assigned | Task | ||
---|---|---|---|---|---|
Open | None | T288287 Remove IE11 from Basic support ("Grade C") | |||
Resolved | ldelench_wmf | T303325 IE11 User and Traffic Analysis |
Event Timeline
MediaWiki's IE11 policy page references reports from 2021 about page-views for top IE11 countries. They are: Pakistan (11.8%), Iran (6.4%), Japan (4.2%), South Korea (5.1%).
I revisited the reports with @Mayakp.wiki (thank you for your help!) and then ran some more in Superset. Would like share findings and propose an analysis.
Findings:
- The analytics team currently disregards traffic from Pakistan, Afghanistan and Iran due to bots and other anomalies, so unfortunately we should remove them from IE11 consideration as well
- Excluding the countries above, here are the top 10 countries with IE11 traffic
- In the above report, South Korea's traffic appears to begin in 2019, but it was formerly reported as Republic of Korea (the traffic does not swap over 1:1, but the same downward trend is apparent)
Questions:
The following reports were pretty strange, what do we think about this?
- Japan report: IE11 traffic in Japan is reported higher than any other browser except Safari 14
- Canada report: @Mayakp.wiki shared that IE11 traffic is high in Canada, but IE6, 7 and 8 also made a comeback in late 2021, surpassing IE11
I would request to assign this briefly to @Mayakp.wiki since she had mentioned that such a spike could be attributed to bots. Would love to hear from the team on the above because IE11 traffic does seem a bit bloated, especially in the report for Japan. Please let me know if I ran this incorrectly. Thanks!
@STHart : Ive shared some notes in this document. pls let me know if you have any questions.
Thank you @Mayakp.wiki! I have a follow-up question - what do you think about this Superset pull vs. browser popularity on Statcounter? Superset shows IE11 traffic as higher than Edge traffic, which doesn't align with overall use. Do you know of a reason why we'd expect to see more IE11 users than Edge users in Japan?
@STHart : I can take a look at it this week, as I have almost wrapped up Tuning Session metrics.
@STHart : I filtered for Desktop traffic to be consistent when comparing with Statcounter. Looking at Overall Edge vs IE traffic, I can see that Edge has higher traffic than IE. The latest versions of Edge definitely have higher traffic than IE11 : Chart
I think it makes sense to look at overall browser family instead of looking more closely to browser major versions, especially since IE doesn't have have a lot of versions compare to other browsers that make more frequent updates.
I believe that the reason Superset showed Edge17 and Edge18 in the top 10 browser versions in the previous chart was because we are looking at traffic from 2018 and the newer versions of Edge, have comparatively lesser traffic than older versions of Edge.
hi @STH, could you please post a summary of the conclusions for this task? also, are the notes in the document linked in T303325#7833753 something that can be made public? thanks!
@ldelench_wmf : Can you please let us know if there is any further support needed from Product Analytics on this task?
Hey @Mayakp.wiki , I'm very sorry for the long delay in responding!
@CCiufo-WMF I'm going to close this task as the initial analysis was completed long ago, and the discussion notes were also made public to anyone with the link. We should submit a new request or sign up for a Product Analytics https://www.mediawiki.org/wiki/Product_Analytics/Consultation_Hours if we need additional support.