mozilla - becoming data driven
DESCRIPTION
Slides from "Becoming Data Driven" session at Mozilla all hands meeting in April 2009.TRANSCRIPT
Becoming a Data Driven Organization
Ken Kovash
22
Agenda
• The “Analytics Umbrella” at Mozilla
• What we’ve done so far
• The Role of our Analytics Function
• What can we do for you?
33
The “Analytics Umbrella” at Mozilla
How can a greater reliance on data driven approaches
enable Mozilla to have a greater impact?
mobile adoption
download experience
website experience
mobile performance
retention experience
product experience
new add-on launches
website experience
website experience
live chat experience
marketing experience
download experience
installation experience
product experience
retention experience
uninstall experience
website experience
l10n
messaging
PR
marketing experience
download experience
installation experience
product experience
js performance
reporter (site compatibility)
crashes
bugzilla
retention experience
uninstall experience
website experience
bugzilla
days vulnerable
PR
new product adoption
test pilot
monetized value of a new user
projecting future sustainability
SFX experience
product launches
campus rep experience
Community marketing
44
Case Study #1
55
Case Study #1
• Our recent Installer Feedback initiative is a good example
• We took a specific area – Fx installation experience – and based
on previous analysis (100K daily users complete download, but
don’t install), we sought to tackle the business problem
• A single day experiment is leading to a couple product changes
(will be implemented for 3.5)
• Resulting impact will be approx. 2,000,000 happier Firefox
users annually (users we would have otherwise lost)
66
Case Study #2
77
Case Study #2
• Fraudulent Firefox related advertisements – what is the size, scope,
and impact (if any) of the problem?
• We took a specific area – Search Engines, which deliver our
“first impression” to 40% of new Fx users – and based on
analysis, estimated that nearly 7 million new Firefox users
(annually) see a fraudulent ad and click through to that
fraudulent site
• We worked with search firms and with our legal team to
determine best approaches for addressing this problem
• Legal currently has a quarterly goal around a systematic
solution; we should be able to successfully make a dent in
improving the user experience for millions of new users
88
So, we’re cool, right?
99
The Role of our Analytics Function
• We’ve seen that the simplest (and often non-sexy)
things can have the biggest impact
• But, we’ve generally seen little leverage
• The previous case studies were based on me
asking myself questions
• Part of this is my fault
• We need to be more proactive
1010
The Role of our Analytics Function
• Interacting with our team…
• Think of us as internal consultants
• lean on us when making a decision and early on
in your initiatives
• i.e., not just after the fact
1111
What can we do for you?
• We’d love to hear from you
• What problem could we solve?
What business question could we answer?
Where can Mozilla be having a greater impact?(these questions can apply to your function or another function)
• Can we help Jay optimize the experience for mobile
users on the new mobile site?
• Can we help Seth understand adoption patterns of
different locales, and what do those patterns tell us about
a language version’s long-term success?
• There are a thousand other questions…