Yesterday we gathered around a virtual fire and shared some things we found
during the past week. Written version first, acted version after:

Phil, like the rest of us, has found that he often wonders if a certain
feature is available in production yet. Instead of just grumbling in IRC,
he made a thing! He calls it FxHey <https://philbooth.me/fxa>, and it
regularly* fetches the version information of our deployed servers*, and
displays it in a fanciful way. Bonus points for source code:
https://github.com/philbooth/fxhey/

Vijay showed us a screenshot of our more *detailed metrics about our
bounced emails*. We can now start to see what locales and what popular
email providers are bouncing more. If you want to see for yourself, here's
the (Mozilla-only) dashboard:
https://kibana-fxa-us-west-2.prod.mozaws.net/#/dashboard/elasticsearch/Auth%20Bouncer%20Logs

Alex was really excited at all the different ways he could *visualize our
metrics in Amplitude*. We've been exploring different options to allow us
to better understand how our users use Firefox Accounts, and Amplitude is
looking like a winner. While we can currently get useful information from
Re:Dash, it requires the Seeker of Information to write SQL, and the
queries can take quite a while to run. Amplitude has a UI that allows
building similar, and in many ways better, queries of our data, and it runs
blazingly fast. It was just a few quicks to build different funnels, and a
particularly nice view was the "Pathfinder": this view allows us to see the
order of events that various users take before reaching the event we're
interested in.

Vlad and Vijay noticed that when creating a new account with an email
address that had been used before, but not verified, the *user would end up
with verification emails with different verification codes*. If the user
then rummages through their email and notices the older verification code,
using it will cause a weird user experience. Their account will be
verified, but we won't have logged them in to the most recent session they
tried to create. There's not end result here, just that this weirdness was
noticed, and fixing it may help with our sign in rate for users who tried
but never finished before.

If you want to see all the action in 4K, well, too bad. If internet video
quality is more up your alley, here's the recording:
https://vreplay.mozilla.com/replay/showRecordDetails.html?recId=4273
_______________________________________________
Dev-fxacct mailing list
[email protected]
https://mail.mozilla.org/listinfo/dev-fxacct

Reply via email to