Mobile Apps Unlocked Vegas 2016
Thursday, May 5
Retention Breakout Session — 3:30pm - 3:50pm
By now, we hope you have notepad’s full of ideas and new thoughts. If for some reason you have not hit your new things to try quota, don’t worry, our closing session will take you over the top featuring first-hand learnings from Life360's non-stop product experimentation engine.
Andreas Gross, Founding Team / Product @ Life360
3. Growth Team
/groTH tem/ n.
A group within the company that develops the product
experience in a way that creates more activated users
! Pre Product/Market fit: build the growth engine
! Post Product/Market fit: optimize the engine (make it bigger)
! Any time: generate hacks, non-obvious ways to reach your goals
Definition
8. Test yourself to product market fit
What type of product are our users interested in?
How can we reach this users the best way?
Pour Gas on it
WE HAD TO EXPERIMENT WITH A LOT OF THINGS
9. How do you build what users actually want?
Child ID
Dog TrackerFamily Locator
Lost Phone
Ads Landing Pages
14. Scalable growth through testing
Lather, Rinse, Repeat. Continuous Improvement
Evaluate
Results
Test
Design
Develop
Hypothesis
Implement
Leanings
15. Developing hypotheses
How to Develop Hypotheses
1. Leverage the team, institutional knowledge,
understand important “Themes”
2. Qualitative studies, surveys
3. Quantitative analysis
16. Test Design
Test
Design
How to Design Tests Effectively
1. Determine right testing format
2. Optimize test to avoid overinvesting
3. Execute quickly, full instrumentation for analysis
17. Evaluate Results
How to Evaluate Results
1. Analyze core metrics impact of each variant
2. Ensure all numbers are statistically significant
3. Understand long term impacts and side effects
Evaluate
Results
18. Implement Learnings
How to Implement Learnings
1. Understand what caused the change
2. Determine how this learnings applies to product aspects
3. Promote visibility of learnings/takeaways
Implement
Leanings
19. Qualitative Tests Painted Door Tests Feature Tests
INVESTMENT COSTLOW HIGH
Painted
Door
Top of
Funnel Test
Full
FeatureMany Some Few
Types of tests avoid overinvesting
20. Qualitative Tests
Purpose
Determine interest in product via conversations
with your users and potential users
Key
Test as many ideas as possible but keep in mind
that people want to please you
Example
Emergency Contacts
ETA Sharing
Carpooling
Qualitative Tests
21. Feature Tests
Purpose
Building a functional feature and see how users
interact
Key
Don’t build the full feature, start at the top of the
funnel and work your way down
Example
Emergency Contacts
Recommender
FUE
Add EC
Send EC
Accept EC
Create ACT
(Partial) Feature Tests
24. Test Design - Explicit Control Variant
Allocation
Design variants to have an additional control group
➔ Ensures control group meets test user criteria
➔ Reduce risk of variant assignment issues
25. Test Design - Explicit Control Variant
Allocation
All Users
26. Test Design - Explicit Control Variant
Allocation
Eligible for experiment
Ineligible for
experiment
27. Test Design - Explicit Control Variant
Allocation
Eligible but excluded from
experiment
Ineligible for
experiment
Included in experiment
28. Test Design - Explicit Control Variant
Allocation
Treatment 1
Eligible but excluded from
experiment
Ineligible for
experiment
Control Variant 1
Variant 2
29. Test Design - Explicit Control Variant
Allocation
Always consider keeping long-term holdout
groups
➔ Understand potential impacts that may not be seen
early in the user lifecycle
➔ Enables measurement of how impacts rollup into
product growth