Closed Bug 1128734 Opened 10 years ago Closed 10 years ago

Create optimizely experiment /firefox/new/ A/B test

Categories

(www.mozilla.org :: Analytics, defect)

Product:
www.mozilla.org
Component:
Analytics
Version:
Production
Platform:
x86
macOS
Type:
defect
Priority:
Not set
Severity:
normal

Tracking

(Not tracked)

Status:
RESOLVED FIXED

People

(Reporter: cmore, Assigned: cmore)

References

Details

(Whiteboard: [kb=1652891] )

This optimizely experiment will sit in the head of this page: https://www.mozilla.org/en-US/firefox/35.0.1/firstrun/ Segment: * en-US visitors * Windows OS * Firefox users Variations: 5 (even split) 1: Redirect to /firstrun/1/ 2: Redirect to /firstrun/2/a/ 3: Redirect to /firstrun/2/b/ 4: Redirect to /firstrun/3/a/ 5: Redirect to /firstrun/3/b/ Sample: 10% (of visitors to the URL above) Goal: Interaction Time period: Feb 12th to Feb 24th Google Analytics integration: yes
Assignee: nobody → chrismore.bugzilla
No longer depends on: 1128708
Blocks: 1120706
Summary: Create optimizely experiment for /firstrun/ A/B test → Create optimizely experiment /firefox/new/ A/B test
This optimizely experiment will sit in the head of this page: https://www.mozilla.org/en-US/firefox/new/ Segment: * en-US visitors * Windows OS * non-Firefox users Variations: 3 (even split) Change download link to: 1) https://download.mozilla.org/?product=firefox-35.0.1-SSL-f34&os=win&lang=en-US 2) https://download.mozilla.org/?product=firefox-35.0.1-SSL-f35&os=win&lang=en-US 3) https://download.mozilla.org/?product=firefox-35.0.1-SSL-f36&os=win&lang=en-US Sample: 10% (of visitors to the URL above) Goal: Interaction Time period: Feb 12th to Feb 24th Google Analytics integration: yes
Whiteboard: [kb=1652891]
I set the goal to download button, but they should be even among all tests because the download page is identical among variations. The real goal is first run interactions and usage that we will measure with GA and FHR.
Status: NEW → RESOLVED
Closed: 10 years ago
Resolution: --- → FIXED
I've reduced the sample to 5% of en-US windows users for the 12 days of the test. We will be monitoring the test and if any problems are discovered, we will shut off the test immediately.
I've enabled the experiment and also adjusted the targeting to be: Windows non-Firefox en-US from organic search I added the organic search, because those users have the cleanest data and will remove some noise between the control and variations.
You need to log in before you can comment on or make changes to this bug.