Steph and Olga work out what the rollout plan for the donate A/B test should be, then run it by the SDS folks
The approved rollout plan is:
Stage 0 - code is rolled out to beta/prod and by default nobody is in the experiment
Stage 1 - Experiment is created in xLab - Testwiki 5/5, English Wikipedia 0/0
Stage 2 - English Wikipedia .05/.05
| Status | Subtype | Assigned | Task | ||
|---|---|---|---|---|---|
| Resolved | SToyofuku-WMF | T395716 [EPIC] Logged-out user A/B test of Donation entry point improvements | |||
| Resolved | ovasileva | T397096 Formalize rollout plan |
My vision is this will be a zero pointer, but reflect the work Olga and I will be doing behind the scenes to figure out how we're going to safely test all paths before launching this experiment
Proposed rollout plan discussed with Olga:
Stage 0 - code is rolled out to beta/prod and nobody is in the experiment, experiment is created in xLab with testwiki at 0% and enwiki at 0%
Stage .5 - Beta 50/50
Stage 1 - Testwiki 50/50, English Wikipedia 0/0
Stage 2 - English Wikipedia target/target <- still need to get this number
Next steps finalize the target for enwiki and double check with experiment platform team about xLab capabilities
Newest plan proposed:
Stage 0 - code is rolled out to beta/prod and by default nobody is in the experiment
Stage .5 - Beta 10/10
Stage 1 - Experiment is created in xLab - Testwiki 10/10, English Wikipedia 0/0
Stage 2 - English Wikipedia .05/.05
I misunderstood the 10% figure:
Stage 0 - code is rolled out to beta/prod and by default nobody is in the experiment
Stage .5 - Beta 5/5
Stage 1 - Experiment is created in xLab - Testwiki 5/5, English Wikipedia 0/0
Stage 2 - English Wikipedia .05/.05
Will double check this before finalizing, then we can add dates when we get closer to finishing the development tasks