acceptance criteria
Set up a/b test for new print styles
notes
Whether we decide to do the test will depend on the outcome of the initial data from T169730: Define and implement instrumentation for printing on desktop web
Description
Description
| Status | Subtype | Assigned | Task | ||
|---|---|---|---|---|---|
| Resolved | Jdlrobson | T154965 [EPIC] Print styles - desktop | |||
| Resolved | ovasileva | T169732 Deploy new desktop print styles on all projects | |||
| Invalid | ovasileva | T169731 Set up print styles a/b test on all projects | |||
| Resolved | • Tbayer | T169730 Define and implement instrumentation for printing on desktop web | |||
| Resolved | phuedx | T171162 [Spike] Answer open questions on instrumentation for printing on desktop web |
Event Timeline
Comment Actions
There is no way to do a 50% roll out on desktop. Those roll puts only work for JSonly solution, thus we need to think about this.
Comment Actions
T169730: Define and implement instrumentation for printing on desktop web now reflects only the instrumentation portion. After the instrumentation is completed and we have established a baseline for the number of users who print multiple times, we will reconsider on whether an a/b test is necessary. Leaving this as stalled for now.
Comment Actions
In T169730#3446799, @Tbayer proposed a potential A/B test design that would negatively affect performance for only 1% of all users, though it seems like it would be limited to users who click the "printable version" link.
Comment Actions
So in the end are we going to do this @ovasileva ? Do we need this placeholder? I'm a little confused.
Comment Actions
@Jdlrobson - most likely not, but I'd like to keep it open until we get a look at the initial data from the printing instrumentation. We can most likely close after that.